regional haze 2nd implementation period four-factor …

REGIONAL HAZE 2ND IMPLEMENTATION PERIOD FOUR-FACTOR ANALYSIS

Ash Grove Cement > Montana City, MT

Montana City Four-Factor Analysis

Prepared By:

Janet Scheier – Managing Consultant Arron Heinerikson, Regional Director

TRINITY CONSULTANTS

9777 Ridge Dr. Suite 380

Lenexa, KS 66219 (913) 894-4500

September 2019

Project 191701.0158

Environmental solutions delivered uncommonly well

Ash Grove Cement Company | Regional Haze 2nd Implementation Period Four-Factor Analysis Trinity Consultants i

TABLE OF CONTENTS

1. EXECUTIVE SUMMARY 1-1

2. INTRODUCTION AND BACKGROUND 2-1

3. SO2 BART EVALUATION 3-1 3.1. Historical Evaluations ............................................................................................................................................. 3-1

3.1.1. 2007 Five Factor Analysis ................................................................................................................................................. 3-1 3.1.2. 2011 Submittal ...................................................................................................................................................................... 3-1

3.2. Update to Montana City Facility and Baseline Emission Rates ................................................................ 3-1 3.3. Additional Analysis .................................................................................................................................................. 3-2

3.3.1. Step 1. Identification of Available Retrofit SO2 Reduction Technologies........................................................ 3-3 3.3.1.1. Fuel Substitution ................................................................................................................................................................. 3-3 3.3.1.2. Wet Scrubbing ...................................................................................................................................................................... 3-3 3.3.1.3. Semi-Dry Scrubbing ........................................................................................................................................................... 3-3

3.3.2. Step 2. Eliminate Technically Infeasible SO2 Control Technologies ................................................................. 3-3 3.3.2.1. Fuel Substitution ................................................................................................................................................................. 3-4 3.3.2.2. Wet Scrubbing ...................................................................................................................................................................... 3-4 3.3.2.3. Semi-Dry Scrubbing ........................................................................................................................................................... 3-4

3.3.3. Step 3. Evaluate Control Effectiveness of Remaining Control Technologies ................................................. 3-4 3.3.4. Step 4. Evaluate Impacts and Document Results .................................................................................................... 3-5

4. NOX BART ANALYSIS 4-1 4.1. Historical Evaluation ............................................................................................................................................... 4-1 4.2. Update to Montana City Facility and Baseline Emission Rates ................................................................ 4-1 4.3. Additional Analysis .................................................................................................................................................. 4-2

4.3.1. Step 1. Identification of Available Retrofit NOx Reduction Technologies ...................................................... 4-2 4.3.1.1. Low NOx Burners (LNB) .................................................................................................................................................. 4-3 4.3.1.2. Selective Catalytic Reduction ........................................................................................................................................ 4-3 4.3.1.3. Selective Non-Catalytic Reduction .............................................................................................................................. 4-4

4.3.2. Step 2. Eliminate Technically Infeasible NOx Control Technologies ................................................................ 4-4 4.3.2.1. Low NOx Burners (LNB) .................................................................................................................................................. 4-4 4.3.2.2. Selective Catalytic Reduction ........................................................................................................................................ 4-4 4.3.2.3. Selective Non-Catalytic Reduction .............................................................................................................................. 4-6

4.3.3. Step 3. Evaluate Control Effectiveness of Remaining Control Technologies ................................................. 4-6 4.3.4. Step 4. Evaluation of Impacts For Feasible NOx Controls .................................................................................... 4-6

4.4. NOx Conclusion .......................................................................................................................................................... 4-7

APPENDIX A: 2007 ANALYSIS A-1

APPENDIX B: 2011 RESPONSE B-1

Ash Grove Cement Company | Regional Haze 2nd Implementation Period Four-Factor Analysis Trinity Consultants ii

LIST OF TABLES

Table 3-1. Evolution of SO2 Baseline Emission Rates 3-2

Table 3-2. Available SO2 Control Technologies for Montana City Kiln 3-3

Table 3-3. Ranking of SO2 Control Technologies by Effectiveness 3-4

Table 4-1. Evolution of NOx Baseline Emission Rates 4-2

Table 4-2. Available NOx Control Technologies for Montana City Kiln 4-3

Table 4-3. Ranking of NOx Control Technologies by Effectiveness 4-6

Ash Grove Cement Company | Regional Haze 2nd Implementation Period Four-Factor Analysis Trinity Consultants 1-1

1. EXECUTIVE SUMMARY

This report documents the results of a four-factor analysis of the cement kiln at Ash Grove Cement Company’s (AGC’s) Montana City facility. In 2007, AGC submitted a BART Five-Factor analysis for the kiln and clinker cooler, and in 2011, AGC submitted a response to a request for additional information regarding the Montana City BART determination.

In April 2019, AGC received a letter from the Montana Department of Environmental Quality (MDEQ) requesting assistance in developing information for a reasonable further progress analysis.

This report addresses any changes made, or updates to the conclusions from, the 2007 and 2011 submittals which are included in Appendices A and B. Using guidelines from EPA for completing the four-step analysis, AGC has determined BART for SO2 and NOx by:

1. Identifying all available retrofit control technologies; 2. Eliminating technically infeasible control technologies; 3. Evaluating the control effectiveness of the remaining control technologies; and 4. Evaluating impacts and document results

Based on the results, AGC believes that reasonable progress compliant controls are already in place as follows:

Long Wet Kiln System (Kiln):

SO2 – AGC installed a semi-dry scrubber in 2012. Baseline SO2 emissions reduced from 981 tpy in 2007 to 101 tpy in 2017/2018. Additional SO2 controls would provide little visibility improvement and require significant expenditures. The facility is limited to 2.0 lb SO2/ton clinker on a 30-day rolling average basis.

NOx – AGC installed a direct-fired low-NOx burner and Selective Non-Catalytic Reduction in September 2014. The baseline emissions have dropped from 1,759 tpy in 2007 to 809 tpy in 2017/2018. The facility is limited to 7.5 lb NOx/ton clinker on a 30-day rolling average basis.


2. INTRODUCTION AND BACKGROUND

Ash Grove’s Montana City plant utilizes a long wet kiln system (kiln) that has been in operation since 1963. The kiln currently employs a baghouse for PM10 control, good combustion practices/low NOx burner and SNCR for NOx control, and semi-dry scrubbing for SO2 control. PM10 emissions from the clinker cooler are also controlled by a baghouse.

In the 1977 amendments to the Clean Air Act (CAA), Congress set a national goal to restore national parks and wilderness areas to natural conditions by preventing any future, and remedying any existing, man-made visibility impairment. On July 1, 1999, the U.S. EPA published the final Regional Haze Rule (RHR). The objective of the RHR is to restore visibility to natural conditions in 156 specific areas across with United States, known as Class I areas. The Clean Air Act defines Class I areas as certain national parks (over 6000 acres), wilderness areas (over 5000 acres), national memorial parks (over 5000 acres), and international parks that were in existence on August 7, 1977.

The RHR requires States to set goals that provide for reasonable progress towards achieving natural visibility conditions for each Class I area in their state. In establishing a reasonable progress goal for a Class I area, the state must:

(A) Consider the costs of compliance, the time necessary for compliance, the energy and non-air quality environmental impacts of compliance, and the remaining useful life of any potentially affected sources, and include a demonstration showing how these factors were taken into consideration in selecting the goal. 40 CFR 51. 308(d)(1)(i)(A).

(B) Analyze and determine the rate of progress needed to attain natural visibility conditions by the year 2064. To calculate this rate of progress, the State must compare baseline visibility conditions to natural visibility conditions in the mandatory Federal Class I area and determine the uniform rate of visibility improvement (measured in deciviews) that would need to be maintained during each implementation period in order to attain natural visibility conditions by 2064. In establishing the reasonable progress goal, the State must consider the uniform rate of improvement in visibility and the emission reduction. 40 CFR 51. 308(d)(1)(i)(B).

In April 2019, AGC received a letter from the MDEQ requesting assistance in developing information for the reasonable further progress analysis.

The information presented in this report considers the following four factors for the emission reductions:

Factor 1. Costs of compliance Factor 2. Time necessary for compliance Factor 3. Energy and non-air quality environmental impacts of compliance Factor 4. Remaining useful life of the kilns

Factors 1 and 3 of the four factors that are listed above were considered by conducting a step-wise review of emission reduction options in a top-down fashion similar to the top-down approach that is included in the EPA


RHR guidelines1 for conducting a review of Best Available Retrofit Technology (BART) for a unit. These steps are as follows:

Step 1. Identify all available retrofit control technologies Step 2. Eliminate technically infeasible control technologies Step 3. Evaluate the control effectiveness of remaining control technologies Step 4. Evaluate impacts and document the results

Factor 4 is also addressed in the step-wise review of the emission reduction options, primarily in the context of the costing of emission reduction options, if any, and whether any capitalization of expenses would be impacted by limited equipment life. Once the step-wise review of reduction options was completed, a review of the timing of the emission reductions is provided to satisfy Factor 2 of the four factors.

1 The BART provisions were published as amendments to the EPA’s RHR in 40 CFR Part 51, Section 308 on July 5, 2005.


3. SO2 BART EVALUATION

Sulfur, in the form of metallic sulfides (pyrite), sulfate, or organosulfur compounds, is often found in the raw materials used to manufacture cement and in the solid and liquid fuels burned in cement kilns.2 The raw materials and fuels for the Montana City plant are no exception. Sulfur dioxide can be generated by the oxidation of sulfur compounds in the raw materials and fuels during operation of the pyroprocess. Constituents found in fuels, raw materials, and in-process materials, such as the alkali metals (sodium and potassium), calcium carbonate, and calcium oxide often react with SO2 within the pyroprocess to limit emissions of SO2 as much of the sulfur leaves the process in the principle product of the kiln system called clinker. As identified in 2007, the kiln is the only BART source at Montana City which emits SO2; thus an SO2 BART evaluation was performed only for the kiln.

3.1. HISTORICAL EVALUATIONS

3.1.1. 2007 Five Factor Analysis

In 2007, AGC submitted a Five Factor Analysis that analyzed four possible retrofit technologies:

Fuel Substitution Raw Material Substitution Lime Spray Dryer (Semi-Wet or Semi-Dry scrubbing) Wet Lime Scrubbing

The 2007 analysis concluded that of these technologies, only Wet Lime Scrubbing and Fuel Substitution were technically feasible (AGC later determined that Wet Lime Scrubbing was not technically feasible and that a semi-dry scrubber was another technically feasible control option). Appendix A contains the 2007 Five Factor Analysis. The visibility analysis performed demonstrated that installation of control for SO2 resulted in little improvement of visibility due to the low contribution of sulfates compared to nitrates on existing visibility. AGC proposed limiting the kiln to the baseline levels of SO2 existing at that time (981 tpy) as BART.

3.1.2. 2011 Submittal

In 2011, EPA asked AGC to support the statement that wet scrubbing was not technically feasible and to submit a 5 factor analysis of dry scrubbing techniques. AGC submitted a response to the inquiry on October 5, 2011 (included in Appendix B). AGC stands by the data submitted regarding wet scrubbing and contends that wet scrubbing is still not technically feasible for the reasons described in the 2011 submittal.

3.2. UPDATE TO MONTANA CITY FACILITY AND BASELINE EMISSION RATES In 2012, AGC installed a semi-dry scrubber, which reduced baseline emissions from a 2010 baseline emission rate of 612 tpy SO2 (see 2011 submittal in Appendix B for detailed calculations) to a future year 2028 estimate of 121 tpy SO2. Table 1 shows the baseline SO2 emission rate evolution since the initial 2007 submittal.

2 Miller, F. MacGregor and Hawkins, Garth J., ”Formation and Emission of Sulfur Dioxide from the Portland Cement Industry”, Proceedings of the Air and Waste Management Association, June 18-22, 2000.


Table 3-1. Evolution of SO2 Baseline Emission Rates

Year

Baseline Annual

SO2 Emissions

(tpy)

Change from 2007

Baseline (tpy)

Notes

2007 981 N/A Existing level of SO2emissions in 2007, calculated by multiplying average emission rate during 2006 (254 lb/hr) by annual operating hours (7,740).

2011 612 -369 In 2007, the SO2 CEMS had only been operating a short time and very little was known about the variability of the SO2 emission rate. Between 2007 and 2011, plant management employed the CEMS to adjust kiln operation such that the SO2 emission rate decreased considerably. The 2011 baseline SO2 emission rate was calculated using the average emission rate during 2010 of 147 lb/hr and a 95 percent availability (8322 hours).

2017/2018 101 -880 AGC installed a semi-dry scrubber in 2012 and further worked with kiln operation to lower SO2 emissions. The 2017/2018 average annual emission rate is provided.

2028 (future estimate)

121 -860 Emissions estimate projected based on 2018 emission rate (0.8 lb/ton clinker) and projected 2028 clinker production (302,000 tons).

3.3. ADDITIONAL ANALYSIS As described in Section 2, Factors 1 and 3 of the four-factor analysis were considered by conducting a step-wise review of emission reduction options in a top-down fashion. The steps are as follows:


Factor 4 is also addressed in the step-wise review of the emission reduction options, primarily in the context of the costing of emission reduction options, if any, and whether any capitalization of expenses would be impacted by a limited equipment life. This section presents the step-wise review of reduction options for SO2. Following the step-wise review of the reduction options for SO2 is a review of the timing of the emission reductions to satisfy Factor 2 of the four factors.

In the original 2007 SO2 BART Evaluation for the kiln, AGC identified four possible retrofit technologies for evaluation. In the 2007 analysis, Raw Material Substitution was removed from consideration since material substitution would result in negligible SO2 reduction. Therefore Raw Material Substitution is not considered in this analysis. Please refer to the 2007 SO2 BART Evaluation in Appendix A for further information.


3.3.1. Step 1. Identification of Available Retrofit SO2 Reduction Technologies

Sulfur dioxide, SO2, is generated during fuel combustion in a cement kiln, as the sulfur in the fuel is oxidized by oxygen in the combustion air. Sulfur in the raw material can also contribute to a kiln’s SO2 emissions.

Step 1 of the top-down control review is to identify available retrofit reduction options for SO2. The available SO2

retrofit control technologies for the Montana City kiln are summarized in Table 3-2. The retrofit controls include both add-on controls that eliminate SO2 after it is formed and switching to lower sulfur fuel that reduces the formation of SO2.

Table 3-2. Available SO2 Control Technologies for Montana City Kiln

SO2 Control Technologies Fuel Substitution

Wet Scrubbing Semi-Dry Scrubbing

3.3.1.1. Fuel Substitution

Fuels that can be considered for the kiln must have sufficient heat content, be dependable and readily available locally in significant quantities to not disrupt continuous production. In addition, they must not adversely affect product quality. Currently, the fuels that the plant is permitted to use, and that are available in continuous quantities, are coal and coke. The ratios of coal/coke can be optimized to minimize SO2 emissions. Alternative lower-sulfur fuel that can be considered is natural gas.

3.3.1.2. Wet Scrubbing

A wet scrubber is a tailpipe technology that may be installed downstream of the kiln. In a typical wet scrubber, the flue gas flows upward through a reactor vessel that has an alkaline reagent flowing down from the top. The scrubber mixes the flue gas and alkaline reagent using a series of spray nozzles to distribute the reagent across the scrubber vessel. The alkaline reagent, often a calcium compound, reacts with the SO2 in the flue gas to form calcium sulfite and/or calcium sulfate that is removed with the scrubber sludge and disposed. Most wet scrubber systems used forced oxidation to assure that only calcium sulfate sludge is produced.

3.3.1.3. Semi-Dry Scrubbing

This technology is considered a semi-wet or semi-dry control technology. A scrubber tower is installed prior to the baghouse. Atomized hydrated lime slurry is sprayed into the exhaust flue gas. The lime absorbs the SO2 in the exhaust and is converted to a powdered calcium/sulfur compound. The particulate control device removes the solid reaction products from the gas stream.

3.3.2. Step 2. Eliminate Technically Infeasible SO2 Control Technologies

Step 2 of the top-down control review is to eliminate technically infeasible SO2 control technologies that were identified in Step 1.


3.3.2.1. Fuel Substitution

In the 2007 analysis, AGC discussed evaluating the coal/coke blend to determine if revising the blend would result in SO2 reductions. Between 2007 and 2011, AGC reduced baseline emissions by 369 tpy (see Table 3-1) through blend revision. AGC continues to evaluate fuel blends in an effort to reduce SO2.

Natural gas can also be considered as a technically feasible replacement for coal/coke as the primary fuel source at this facility, and can be evaluated further. For natural gas to be a technically feasible option, the supply of natural gas must be reliable on a continuous basis. While the Montana City facility uses natural gas for startup, the facility has been curtailed by the natural gas supplier the last two winters. And, the supplier cannot guarantee a continuous (free from curtailment) supply of natural gas in the future. Further, AGC has experienced extended downtime at another facility as a result of being reliant on natural gas. AGC’s Seattle facility uses natural gas as their primary fuel. On October 9, 2018, a 36-inch natural gas pipeline ruptured in British Columbia, causing the two main natural gas supply lines to the Seattle area to be shut down. The Seattle facility had to stop production for more than a month while supply was stabilized and routed to more critical infrastructure users, such as electric utilities. Consequently, natural gas is not considered available on a continuous basis, and relying on natural gas to be the sole fuel source for the facility is not feasible.

3.3.2.2. Wet Scrubbing

In the 2011 analysis, AGC evaluated use of a wet scrubber and demonstrated that a wet scrubber was technically infeasible for the Montana City Facility. Please refer to Appendix B and the 2011 discussion. Therefore, Wet Scrubbing is deemed technically infeasible and is removed from consideration.

3.3.2.3. Semi-Dry Scrubbing

AGC already uses this technology, and installed a semi-wet/dry scrubber in 2012. Semi-wet/dry scrubbing is technically feasible and will be considered further.

3.3.3. Step 3. Evaluate Control Effectiveness of Remaining Control Technologies

Step 3 of the top-down control review is to rank the technically feasible options by effectiveness. Table 3-3 presents available and feasible SO2 control technologies for the kiln and their associated reduction efficiencies.

Table 3-3. Ranking of SO2 Control Technologies by Effectiveness

Pollutant

Control

Technology

Potential Reduction Efficiency

(%)

SO2 Semi-Dry Scrubbinga 80 – 90 Fuel Substitutionb 30 - 50

a Semi-dry Scrubber reduction efficiency was estimated based on the actual reduction from 2011 levels to current levels as shown in Table 3-1.

b Fuel substitution reduction efficiency was estimated based on the actual reduction in SO2 emissions of approximately 40% from 2007 to 2011 as shown in Table 3-1.


3.3.4. Step 4. Evaluate Impacts and Document Results

Step 4 of the top-down control review is the impact analysis. While the impact analysis considers the cost of compliance, energy impacts, non-air quality impacts, and the remaining useful life of the source, AGC has installed semi-dry scrubbing which is the control strategy with the greatest level of control.

Therefore, AGC believes that reasonable progress compliant controls are already in place. As shown in Table 3-1, AGC’s SO2 emissions have been reduced by over 860 tpy from the 2007 baseline, and the 2007 visibility analysis showed that reduction in SO2 emissions would have little improvement on visibility. As a result, AGC proposes that the existing levels of SO2 (projected 2028 actuals of 121 tpy SO2) are adequate and the current controls constitute BART for the kiln. Further, Ash Grove does not propose any change to their current limit of 2.0 lb SO2/ton clinker on a 30-day rolling average basis.

Ash Grove Cement Company | Regional Haze 2nd Implementation Period Four-Factor Analysis 4-1 Trinity Consultants

4. NOX BART ANALYSIS

In Portland cement kilns, the NOx that is generated is primarily classified into one of two categories, i.e., thermal NOx or fuel NOx3. Thermal NOx occurs as a result of the high-temperature oxidation of molecular nitrogen present in the combustion air. Fuel NOx is created by the oxidation of nitrogenous compounds present in the fuel. It is also possible for nitrogenous compounds to be present in the raw material feed and become oxidized to form additional NOx referred to as feed NOx. Due to the high flame temperature in the burning zone of the rotary kiln (3400o F), NOx emissions from the kiln tend to be mainly comprised of thermal NOx. Although NOx emissions from cement kilns include both nitrogen oxide (NO) and nitrogen dioxide (NO2), typically, less than 10% of the total NOx in the flue gas is NO2.4 As identified in 2007, the kiln is the only BART source which emits NOx, thus a NOx BART evaluation was performed only for the kiln.

4.1. HISTORICAL EVALUATION In 2007, AGC submitted a Five Factor Analysis that analyzed six possible retrofit technologies:

Low NOx Burner (LNB) Flue Gas Recirculation CKD Insufflation Mid-Kiln Firing of Tires Selective Noncatalytic Reduction (SNCR) Selective Catalytic Reduction (SCR)

The 2007 analysis concluded that of these technologies, only LNB and SNCR (either singularly or in combination) were technically feasible. Appendix A contains the 2007 Five Factor Analysis. The visibility analysis performed demonstrated significant improvement with installation of SNCR and LNB. Therefore, AGC proposed that a direct-fired LNB system with SNCR constituted BART. AGC proposed to comply with a BART emission limit of 227.25 lb/hr on a 30-day rolling basis.

4.2. UPDATE TO MONTANA CITY FACILITY AND BASELINE EMISSION RATES In 2014, AGC installed an SNCR, which along with the direct fire LNB, has reduced baseline emissions to a future year 2028 estimate of 981 tpy NOx from a 2007 baseline emission rate of 1,759 tpy NOx. Table 4-1 shows the comparison of the 2007 baseline NOx emission rate to the current projected 2028 baseline emission rate.

3 NOx Formation and Variability in Portland Cement Kiln Systems, Penta Engineering, December 1998.

4 IBID.


Table 4-1. Evolution of NOx Baseline Emission Rates

Year

Baseline Annual

NOx Emissions

(tpy)

Change from 2007

Baseline (tpy)

Notes

2007 1759 N/A Existing level of NOx emissions in 2007, calculated by multiplying average emission rate during 2006 (454.5 lb/hr) by annual operating hours (7,740).

2018 809 -950 AGC installed an SNCR in 2014, and along with the direct fire LNB, realized greater than a 50% decrease in emissions.

2028 (future estimate)

981 -778 Emissions estimate projected based on 2018 emission rate of 6.5 lb/ton and projected 2028 clinker production of 302,000 tons.

4.3. ADDITIONAL ANALYSIS As described in Section 2, Factors 1 and 3 of the four-factor analysis were considered by conducting a step-wise review of emission reduction options in a top-down fashion. The steps are as follows:


Factor 4 is also addressed in the step-wise review of the emission reduction options, primarily in the context of the costing of emission reduction options and whether any capitalization of expenses would be impacted by a limited equipment life. This section presents the step-wise review of reduction options for NOx. Following the step-wise review of the reduction options for NOx is a review of the timing of the emission reductions to satisfy Factor 2 of the four factors.

In the original 2007 NOx BART Evaluation for the kiln, AGC identified six possible retrofit technologies for evaluation. In the 2007 analysis, flue gas recirculation, cement kiln dust insufflation, and mid-kiln firing of solid fuel (tires), were eliminated from consideration due to factors that still exist for the Montana City facility. Therefore, these NOx reduction strategies are not considered further in this analysis. Please refer to the 2007 NOx BART Evaluation in Appendix A for further information.

4.3.1. Step 1. Identification of Available Retrofit NOx Reduction Technologies

Nitrogen oxides, NOx, are produced during fuel combustion when nitrogen contained in the fuel and combustion air is exposed to high temperatures. The origin of the nitrogen (i.e. fuel vs. combustion air) has led to the use of the terms “thermal” NOX and “fuel” NOx when describing NOx emissions from the combustion of fuel. Thermal NOX emissions are produced when elemental nitrogen in the combustion air is admitted to a high temperature zone and oxidized. Fuel NOx emissions are created during the rapid oxidation of nitrogen compounds contained in the fuel.


Most of the NOx formed within a cement kiln is classified as thermal NOx. Virtually all of the thermal NOX is formed in the region of the flame at the highest temperatures, approximately 3,000 to 3,600 degrees Fahrenheit. A small portion of NOx is formed from nitrogen in the fuel that is liberated and reacts with the oxygen in the combustion air.

Step 1 of the top-down control review is to identify available retrofit reduction options for NOX. The remaining available NOX retrofit control technologies for the Montana City are summarized in Table 4-2 (other control technologies eliminated during the 2007 analysis are not considered).

Table 4-2. Available NOx Control Technologies for Montana City Kiln

NOx Control Technologies Combustion Controls Low NOx Burners (LNB)

Post-Combustion Controls Selective Catalytic Reduction (SCR) Selective Non-Catalytic Reduction (SNCR)

NOX emissions controls, as listed in Table 4-2, can be categorized as combustion or post-combustion controls. Combustion controls reduce the peak flame temperature and excess air in the kiln burner, which minimizes NOX formation. Post-combustion controls such as selective catalytic reduction (SCR) and selective non-catalytic reduction (SNCR), convert NOX in the flue gas to molecular nitrogen and water.

4.3.1.1. Low NOx Burners (LNB)

Low-NOX Burners (LNBs) reduce the amount of NOX initially formed in the flame. The principle of all LNBs is the same: stepwise or staged combustion and localized exhaust gas recirculation (i.e., at the flame). LNBs are designed to reduce flame turbulence, delay fuel/air mixing, and establish fuel-rich zones for initial combustion. The longer, less intense flames reduce thermal NOX formation by lowering flame temperatures. Some of the burner designs produce a low-pressure zone at the burner center by injecting fuel at high velocities along the burner edges. Such a low-pressure zone tends to recirculate hot combustion gas, which is retrieved through an internal reverse flow zone around the extension of the burner centerline. The recirculated combustion gas is deficient in oxygen, thus producing the effect of flue gas recirculation. Reducing the oxygen content of the primary air creates a fuel-rich combustion zone that then generates a reducing atmosphere for combustion. Due to fuel-rich conditions and lack of available oxygen, formation of thermal NOX and fuel NOX are minimized

4.3.1.2. Selective Catalytic Reduction

Selective catalytic reduction (SCR) is an exhaust gas treatment process in which ammonia (NH3) is injected into the exhaust gas upstream of a catalyst bed. On the catalyst surface, NH3 and nitric oxide (NO) or nitrogen dioxide (NO2) react to form diatomic nitrogen and water. The overall chemical reactions can be expressed as follows:

4NO + 4NH3+O2→4N2 + 6H2O

2NO2+4NH3+O2→3N2+6H2O


When operated within the optimum temperature range of 480°F to 800°F, the reaction can result in removal efficiencies between 70 and 90 percent.5 The rate of NOX removal increases with temperature up to a maximum removal rate at a temperature between 700°F and 750°F. As the temperature increases above the optimum temperature, the NOX removal efficiency begins to decrease. The application of SCR is extremely limited in the U.S. cement industry, as only one cement plant has installed SCR for NOx control (in 2015) and the specifics of its installation and use remain confidential.

4.3.1.3. Selective Non-Catalytic Reduction

In SNCR systems, a reagent is injected into the flue gas within an appropriate temperature window. The NOX and reagent (ammonia or urea) react to form nitrogen and water. A typical SNCR system consists of reagent storage, multi-level reagent-injection equipment, and associated control instrumentation. The SNCR reagent storage and handling systems are similar to those for SCR systems. Like SCR, SNCR uses ammonia or a solution of urea to reduce NOx through a similar chemical reaction.

2NO+4NH3+2O2→3N2+6H2O

SNCR requires a higher temperature range than SCR of between 1,600°F and 1,900°F due to the lack of a catalyst to lower the activation energies of the reactions.

4.3.2. Step 2. Eliminate Technically Infeasible NOx Control Technologies

Step 2 of the top-down control review is to eliminate technically infeasible NOx control technologies that were identified in Step 1.

4.3.2.1. Low NOx Burners (LNB)

The AGC Montana City facility installed LNB after the 2007 analysis. This technology currently operates at the Montana City facility.

4.3.2.2. Selective Catalytic Reduction

Efficient operation of the SCR process requires constant exhaust temperatures (usually ± 200°F).6 Fluctuation in exhaust gas temperatures reduces removal efficiency. If the temperature is too low, ammonia slip occurs. Ammonia slip is caused by low reaction rates and results in both higher NOX emissions and appreciable ammonia emissions. If the temperature is too high, oxidation of the NH3 to NO can occur. Also, at higher removal efficiencies (beyond 80 percent), an excess of NH3 is necessary, thereby resulting in some ammonia slip. Other emissions possibly affected by SCR include increased PM emissions (from ammonia salts in a detached plume) and increased SO3 emissions (from oxidation of SO2 on the catalyst). These ammonia, PM, and ammonia salt emissions contribute negatively to visibility impairment in the region—an effect that is directly counter to the goals of the program.

5 Air Pollution Control Cost Manual, Section 4, Chapter 2, Selective Catalytic Reduction, NOx Controls, EPA/452/B-02-001,

Page 2-9 and 2-10. 6 USEPA, Office of Air Quality Planning and Standards. Alternative Control Technologies Document – NOx Emissions from

Cement Manufacturing. EPA-453/R-94-004, Page 2-11


To reduce fouling the catalyst bed with the PM in the exhaust stream, an SCR unit can be located downstream of the particulate matter control device (PMCD). However, due to the low exhaust gas temperature exiting the PMCD (approximately 350 °F); a heat exchanger system would be required to reheat the exhaust stream to the desired reaction temperature range of between 480 °F to 800 °F. The source of heat for the heat exchanger would be the combustion of fuel, with combustion products that would enter the process gas stream and generate additional NOX.7 Therefore, in addition to storage and handling equipment for the ammonia, the required equipment for the SCR system will include a catalytic reactor, heat exchanger and potentially additional NOX control equipment for the emissions associated with the heat exchanger fuel combustion. High dust and clean-side SCR technologies are still highly experimental. A high dust SCR would be installed prior to the dust collectors, where the kiln exhaust temperature is closer to the optimal operating range for an SCR. It requires a larger volume of catalyst than a tail pipe unit, and a mechanism for periodic cleaning of catalyst. A high dust SCR also uses more energy than a tail pipe system due to catalyst cleaning and pressure losses. A clean-side system is similar to a high dust system. However, the SCR is placed downstream of the baghouse. Only one cement kiln in the U. S. is using SCR, and the details of its installation and use remain confidential. While several cement kilns in Europe have installed SCR, the cement industries between Europe and the U.S. differ significantly due to the increased sulfur content found in the processed raw materials in U.S. cement kiln operations. The pyritic sulfur found in raw materials used by U.S. cement plants have high SO3 concentrations that result in high-dust levels and rapid catalyst deactivation. In the presence of calcium oxide and ammonia, SO3 forms calcium sulfate and ammonium bisulfate via the following reactions:

SO3 + CaO → CaSO4

SO3 + NH3 → (NH4)HSO4

Calcium sulfate can deactivate the catalyst, while ammonium bisulfate can plug the catalyst. Catalyst poisoning can also occur through the exposure to sodium, potassium, arsenic trioxide, and calcium sulfate.8 This effect directly and negatively impacts SCR effectiveness for NOX reduction. Dust buildup on the catalyst is influenced by site-specific raw material characteristics present in the facility’s quarry, such as trace contaminants that may produce a stickier particulate than is experienced at sites where the technology is being demonstrated. This buildup is typical of cement kilns, resulting in reduced effectiveness, catalyst cleaning challenges, and increased kiln downtime at significant cost.9

In the EPA’s guidance for regional haze analysis, the term “available,” one of two key qualifiers for technical feasibility in a BART analysis, is clarified with the following statement:

7 The fuel would likely be natural gas supplied at the facility through a pipeline while coal will be excluded, as it would

require an additional dust collector. 8 Air Pollution Control Cost Manual, Section 4, Chapter 2, Selective Catalytic Reduction, NOx Controls, EPA/452/B-02-001, Page 2-6 and 2-7. 9 Preamble to NSPS subpart F, 75 FR 54970.


Consequently, you would not consider technologies in the pilot scale testing stages of development as “available” for the purposes of BART review.

The EPA has also acknowledged, in response to comments made by the Portland Cement Association’s (PCA) comments on the latest edition of the Control Cost Manual, that:

For some industrial applications, such as cement kilns where flue gas composition varies with the raw materials used, a slip stream pilot study can be conducted to determine whether trace elements and dust characteristics of the flue gas are compatible with the selected catalyst.

Based on these conclusions, SCR is not widely available for use with cement kilns, in large part because the site-specificity limits the commercial availability of systems. For this reason, high-dust and clean-side SCR’s are not considered technically feasible for this facility at this time.

4.3.2.3. Selective Non-Catalytic Reduction

The AGC Montana City facility installed SNCR in 2014. This technology currently operates at the Montana City facility.

4.3.3. Step 3. Evaluate Control Effectiveness of Remaining Control Technologies

Step 3 of the top-down control review is to rank the technically feasible options to effectiveness. Table 4-3 presents available and feasible NOX control technologies for the kilns and their associated control efficiencies.

Table 4-3. Ranking of NOx Control Technologies by Effectiveness

Pollutant Control

Technology Potential

Effectiveness NOx SNCR + LNB 6.5 lb/tona

a Current average annual actual emission rate based on 2017/2018 data. The current NOx limit is 7.5 lb/ton.

4.3.4. Step 4. Evaluation of Impacts For Feasible NOx Controls

Step 4 of the top-down control review is the impact analysis. While the impact analysis considers the cost of compliance, energy impacts, non-air quality impacts, and the remaining useful life of the source, AGC has installed the control strategy with the greatest level of control: SNCR + LNB.

Therefore, AGC believes that reasonable progress compliant controls are already in place. As shown in Table 4-1, AGC’s NOx emissions have been reduced by over 778 tpy from the 2007 baseline. As a result, AGC proposes that the existing levels of NOx (projected 2028 actuals of 981 tpy NOx) are adequate and the current controls constitute BART for the kiln.


4.4. NOX CONCLUSION

The AGC Montana City facility currently utilizes Low- NOx burners and SNCR to control NOx emissions. , AGC believes that the current technologies of LNB and SNCR represent BART for NOx. Further, Ash Grove does not propose any change to their current limit of 7.5 lb NOX/ton clinker on a 30-day rolling average basis.

Ash Grove Cement Company | Regional Haze 2nd Implementation Period Four-Factor Analysis A-1 Trinity Consultants

APPENDIX A: 2007 ANALYSIS

BART FIVE FACTOR ANALYSIS ASH GROVE CEMENT MONTANA CITY, MONTANA

VERSION 0

Prepared by:

TRINITY CONSULTANTS 9777 Ridge Drive

Suite 380 Lenexa, Kansas 66219

(913) 894-4500

June 2007

Project 071601.0069

Ash Grove Cement Company i Trinity Consultants BART Analysis

TABLE OF CONTENTS

1. EXECUTIVE SUMMARY ........................................................................................ 1-1

2. INTRODUCTION AND BACKGROUND.................................................................... 2-3

3. BART APPLICABILITY DETERMINATION........................................................... 3-5

4. SO2 BART EVALUATION .................................................................................... 4-1 4.1 IDENTIFICATION OF AVAILABLE RETROFIT SO2 CONTROL TECHNOLOGIES .......... 4-1 4.2 ELIMINATE TECHNICALLY INFEASIBLE SO2 CONTROL TECHNOLOGIES ................ 4-2

4.2.1 FUEL SUBSTITUTION ...........................................................................................4-2 4.2.2 RAW MATERIAL SUBSTITUTION .........................................................................4-2 4.2.3 LIME SPRAY DRYING ..........................................................................................4-3 4.2.4 WET LIME SCRUBBING .......................................................................................4-3

4.3 RANK OF TECHNICALLY FEASIBLE SO2 CONTROL OPTIONS BY EFFECTIVENESS .. 4-4 4.4 EVALUATION OF IMPACTS FOR FEASIBLE SO2 CONTROLS..................................... 4-4

4.4.1 WET LIME SCRUBBING .......................................................................................4-5 4.4.2 FUEL SUBSTITUTION ...........................................................................................4-9

4.5 EVALUATION OF VISIBILITY IMPACT OF FEASIBLE SO2 CONTROLS .................... 4-12 4.6 PROPOSED BART FOR SO2..................................................................................... 4-1

5. NOX BART EVALUATION................................................................................... 5-3 5.1 IDENTIFICATION OF AVAILABLE RETROFIT NOX CONTROL TECHNOLOGIES......... 5-3 5.2 ELIMINATE TECHNICALLY INFEASIBLE NOX CONTROL TECHNOLOGIES ............... 5-4

5.2.1 LOW-NOX BURNER IN THE ROTARY KILN .........................................................5-4 5.2.2 FLUE GAS RECIRCULATION ................................................................................5-6 5.2.3 CEMENT KILN DUST INSUFFLATION ...................................................................5-6 5.2.4 MID-KILN FIRING OF SOLID FUEL (TIRES) WITH MIXING AIR FAN....................5-7 5.2.5 SELECTIVE NONCATALYTIC REDUCTION............................................................5-8 5.2.6 SELECTIVE CATALYTIC REDUCTION...................................................................5-9

5.3 RANK OF TECHNICALLY FEASIBLE NOX CONTROL OPTIONS BY EFFECTIVENESS 5-12 5.4 EVALUATION OF IMPACTS FOR FEASIBLE NOX CONTROLS ................................. 5-13

5.4.1 SNCR AND LNB ...............................................................................................5-13 5.5 EVALUATION OF VISIBILITY IMPACT OF FEASIBLE NOX CONTROLS ................... 5-14 5.6 PROPOSED BART FOR NOX.................................................................................. 5-15

6. PM BART EVALUATION..................................................................................... 6-1

Ash Grove Cement Company ii Trinity Consultants BART Analysis

LIST OF TABLES

TABLE 1-1. VISIBILITY IMPAIRMENT IMPROVEMENT AT GATES OF THE MOUNTAINS WILDERNESS AREA ......................................................................................................................................1-2

TABLE 3-1. SUMMARY OF U.S. EPA REGION 8 BART APPLICABILITY MODELING RESULTS ...........3-6

TABLE 3-2. SUMMARY OF MODELING METHOD DIFFERENCES BETWEEN EPA AND AGC ...............3-7

TABLE 3-3. KILN EMISSION RATES IN BART DATA SUBMITTAL VS. 2006 ANALYZER DATA...........3-8

TABLE 3-4. EXISTING MAXIMUM 24-HOUR SO2, NOX, AND PM10 EMISSIONS (AS HOURLY EQUIVALENTS) .......................................................................................................................3-8

TABLE 3-5. SUMMARY OF EXISTING STACK PARAMETERS ................................................................3-9

TABLE 3-6. EXISTING VISIBILITY IMPAIRMENT ATTRIBUTABLE TO MONTANA CITY PLANT KILN AND CLINKER COOLER...................................................................................................................3-9

TABLE 4-1. EXISTING ACTUAL MAXIMUM 24-HOUR SO2 EMISSION RATES......................................4-1

TABLE 4-2. AVAILABLE SO2 CONTROL TECHNOLOGIES.....................................................................4-2

TABLE 4-3. RANKING OF TECHNICALLY FEASIBLE KILN SO2 CONTROL TECHNOLOGIES BY EFFECTIVENESS ......................................................................................................................4-4

TABLE 4-4. COST ANALYSIS FOR WET LIME SCRUBBING ..................................................................4-6

TABLE 4-4. COST ANALYSIS FOR WET LIME SCRUBBING (CONTINUED)............................................4-7

TABLE 4-4. COST ANALYSIS FOR WET LIME SCRUBBING (CONTINUED)............................................4-8

TABLE 4-5. SUMMARY OF COST EFFECTIVENESS FOR FUEL SUBSTITUTION....................................4-11

TABLE 4-6. SUMMARY OF EMISSION RATES MODELED IN SO2 CONTROL VISIBILITY IMPACT ANALYSIS .............................................................................................................................4-12

TABLE 4-7. SUMMARY OF MODELED IMPACTS FROM SO2 CONTROL VISIBILITY IMPACT ANALYSIS4-1

TABLE 4-8. SUMMARY OF COST EFFECTIVENESS OF SO2 CONTROL TECHNOLOGIES ........................4-1

TABLE 5-1. EXISTING ACTUAL MAXIMUM 24-HOUR NOX EMISSION RATES.....................................5-3

TABLE 5-2. POSSIBLE NOX CONTROL TECHNOLOGIES ......................................................................5-4

TABLE 5-3. RANKING OF TECHNICALLY FEASIBLE KILN NOX CONTROL TECHNOLOGIES BY EFFECTIVENESS ....................................................................................................................5-13

Ash Grove Cement Company iii Trinity Consultants BART Analysis

TABLE 5-4. SUMMARY OF EMISSION RATES MODELED IN NOX CONTROL VISIBILITY IMPACT ANALYSIS .............................................................................................................................5-15

TABLE 5-5. NOX CONTROL VISIBILITY IMPACT ANALYSIS ...............................................................5-15

TABLE 6-1. EXISTING MAXIMUM 24-HOUR PM10 EMISSION RATE....................................................6-1

TABLE 6-2. VAP VISIBILITY IMPAIRMENT CONTRIBUTIONS AT GATES OF THE MOUNTAINS............6-1

Ash Grove Cement Company 1-1 Trinity Consultants BART Analysis

1. EXECUTIVE SUMMARY

This report documents the determination of the Best Available Retrofit Technology (BART) as proposed by Ash Grove Cement Company (AGC) for the Portland cement manufacturing plant located in Montana City, Montana (Montana City plant). There are two emission units at the Montana City plant for which AGC has made a BART determination: the kiln and the clinker cooler. Currently, particulate matter emissions from the kiln are controlled by an electrostatic precipitator. Particulate matter emissions from the clinker cooler are controlled by a baghouse. The Montana City plant has other lesser emitting BART-eligible emissions units, but the negligible visibility impairment attributable to these sources concludes that no additional controls are necessary to satisfy the requirements of the BART rule.1 AGC used the U.S. Environmental Protection Agency’s (EPA’s) guidelines2 in 40 CFR Part 51 to determine BART for the kiln and clinker cooler. Specifically, AGC conducted a five-step analysis to determine BART for SO2, NOX, and PM10 that included the following: 1. Identifying all available retrofit control technologies; 2. Eliminating technically infeasible control technologies; 3. Evaluating the control effectiveness of remaining control technologies; 4. Evaluating impacts and document the results; 5. Evaluating visibility impacts Based on the five-step analysis, AGC proposes the following as BART: Kiln:

• PM10 – AGC proposes that the existing electrostatic precipitator constitutes BART. This control device is the most effective for controlling PM10 from a wet kiln.

• NOX – AGC proposes to comply with a BART emission limit of 227.25 lb/hr on a 30-day

rolling basis by installing and operating a direct-fired low-NOx burner (LNB) and a selective noncatalytic reduction (SNCR) system. Compliance with the emission limit will be demonstrated by continuous emissions monitoring.

• SO2 – AGC proposes that no additional SO2 controls are required for BART compliance.

Additional SO2 controls would provide little visibility improvement and require significant expenditures.

Clinker Cooler:

1 AGC submitted an inventory of all of the BART-eligible emission sources to the Montana Department of

Environmental Quality. EPA Region 8 subsequently evaluated the kiln and clinker cooler to determine the applicability of BART to the Montana City plant. Trinity Consultants conducted two BART applicability visibility modeling analyses for the Montana City plant. One modeling analysis included all of the BART-eligible sources at the plant. The other modeling analysis included the kiln and clinker cooler only. The difference in the modeled visibility impact predicted for the two scenarios was negligible; thus, it was concluded that the contribution of the non-kiln and clinker cooler sources to visibility impairment is negligible, and controlling these sources would not improve any existing visibility impairment.

2 40 CFR 51, Regional Haze Regulations and Guidelines for Best Available Retrofit Technology (BART) Determinations


• PM10 – AGC proposes that the existing baghouse constitutes BART. This control device is the most effective for controlling PM10 from a clinker cooler.

The proposed BART control strategies will result in reductions of the visibility impacts attributable to the Montana City plant. A summary of the visibility improvement at the Gates of the Mountains Class I area based on the existing emission rates and proposed BART emission rates is provided in Table 1-1.

TABLE 1-1. VISIBILITY IMPAIRMENT IMPROVEMENT AT GATES OF THE MOUNTAINS WILDERNESS AREA

98% Impact

(Δdv) Existing 2.874 BART 1.377

Improvement 52.09%


2. INTRODUCTION AND BACKGROUND

On July 1, 1999, the U.S. EPA published the final Regional Haze Rule (RHR). The objective of the RHR is to improve visibility in 156 specific areas across with United States, known as Class I areas. The Clean Air Act defines Class I areas as certain national parks (over 6000 acres), wilderness areas (over 5000 acres), national memorial parks (over 5000 acres), and international parks that were in existence on August 7, 1977. On July 6, 2005, the EPA published amendments to its 1999 RHR, often called the Best Available Retrofit Technology (BART) rule, which included guidance for making source-specific BART determinations. The BART rule defines BART-eligible sources as sources that meet the following criteria:

(1) Have potential emissions of at least 250 tons per year of a visibility-impairing pollutant, (2) Began operation between August 7, 1962 and August 7, 1977, and (3) Are included as one of the 26 listed source categories in the guidance.

A BART-eligible source is subject to BART if the source is “reasonably anticipated to cause or contribute to visibility impairment in any federal mandatory Class I area.” EPA has determined that a source is reasonably anticipated to contribute to visibility impairment if the 98th percentile visibility impacts from the source are greater than 0.5 delta deciviews (∆dv) when compared against a natural background. Air quality modeling is the tool that is used to determine a source’s visibility impacts. Once it is determined that a source is subject to BART, a BART determination must address air pollution control measures for the source. The visibility regulations define BART as follows:

“…an emission limitation based on the degree of reduction achievable through the application of the best system of continuous emission reduction for each pollutant which is emitted by…[a BART-eligible source]. The emission limitation must be established on a case-by-case basis, taking into consideration the technology available, the cost of compliance, the energy and non air quality environmental impacts of compliance, any pollution control equipment in use or in existence at the source, the remaining useful life of the source, and the degree of improvement in visibility which may reasonable be anticipated to result from the use of such technology.

Specifically, the BART rule states that a BART determination should address the following five statutory factors: 1. Existing controls 2. Cost of controls 3. Energy and non-air quality environmental impacts 4. Remaining useful life of the source 5. Degree of visibility improvement as a result of controls Further, the BART rule indicates that the five basic steps in a BART analysis can be summarized as follows: 1. Identify all available retrofit control technologies;


2. Eliminate technically infeasible control technologies; 3. Evaluate the control effectiveness of remaining control technologies; 4. Evaluate impacts and document the results; 5. Evaluate visibility impacts A BART determination should be made for each visibility affecting pollutant (VAP) by following the five steps listed above for each VAP. BART applicability was determined for the Montana City plant based on a combination of an applicability analysis performed by U.S. EPA Region 8 and a refined applicability analysis performed by AGC. Both analyses determined that the kiln and clinker cooler are subject to BART. The details of the applicability determination can be found in Section 3. Subsequently, AGC performed an analysis to determine BART for each VAP for the kiln and clinker cooler. The VAPs emitted by the kiln and clinker cooler include NOx, SO2, and particulate matter with a mass mean diameter smaller than ten microns (PM10) of various forms (filterable coarse particulate matter [PMc], filterable fine particle matter [PMf], elemental carbon [EC], inorganic condensable particulate matter [IOR CPM] as sulfates [SO4], and organic condensable particulate matter [OR CPM] also referred to as secondary organic aerosols [SOA]). The BART determinations for SO2, NOX, and PM10 can be found in Sections 4, 5, and 6, respectively.


3. BART APPLICABILITY DETERMINATION

As stated in Section 2, a BART-eligible source is subject-to-BART if the source is “reasonably anticipated to cause or contribute to visibility impairment in any federal mandatory Class I area.” EPA has determined that a source is reasonably anticipated to contribute to visibility impairment if the 98th percentile of the visibility impacts from the source is greater than 0.5 ∆dv when compared against a natural background. U.S. EPA Region 8 (EPA) conducted air quality modeling to predict the existing visibility impairment attributable to the Montana City plant in the following Class I areas:

• Gates of the Mountains Wilderness Area • Scapegoat Wilderness Area • Anaconda – Pintler Wilderness Area • Bob Marshall Wilderness Area • Mission Mountains Wilderness Area • Selway – Bitterroot Wilderness • Yellowstone National Park • Red Rock Lakes Wilderness Area • Glacier National Park • North Absaroka Wilderness Area • Washakie Wilderness Area • Teton Wilderness Area

Based on this modeling, EPA concluded that the Montana City plant was subject to BART since the 98th percentile of the visibility impacts attributable to the kiln and clinker cooler are greater than 0.5 Δdv when compared against a natural background for one Class I area: Gates of the Mountains Wilderness Area. The results of the applicability modeling are summarized in Table 3-1.


TABLE 3-1. SUMMARY OF U.S. EPA REGION 8 BART APPLICABILITY MODELING RESULTS

Minimum Distance

98th Percentile Visibility Impact for Each Year

(∆ dv)

Overall 98th

Percentile Visibility Impact

Class I Area (km) 2001 2002 2003 (∆ dv)

Gates of the Mountains 30 2.17 1.82 2.52 2.17 Scapegoat 80 0.36 0.42 0.26 0.34 Anaconda - Pintler 113 0.09 0.09 0.07 0.08 Bob Marshall 116 0.39 0.30 0.18 0.30 Mission Mountains WA 162 0.05 0.06 0.04 0.05 Selway-Bitterroot Wilderness 173 0.01 0.01 0.01 0.01 Yellowstone NP 175 0.00 0.01 0.01 0.01 Red Rock Lakes 207 0.00 0.00 0.00 0.00 Glacier NP 223 0.10 0.04 0.04 0.05 North Absaroka Wilderness 228 0.00 0.00 0.00 0.00 Washakie Wilderness 276 0.00 0.00 0.00 0.00 Teton Wilderness 289 0.00 0.00 0.00 0.00

AGC verified EPA’s results by performing a refined modeling analysis for the Class I area located closest to the Montana City plant: Gates of the Mountains Wilderness Area. The modeling methods used by AGC and EPA Region 8 differed slightly and are summarized in Table 3-2.


TABLE 3-2. SUMMARY OF MODELING METHOD DIFFERENCES BETWEEN EPA AND AGC

Processor/Model Parameter Ash Grove (AGC) Modeling Analysis EPA Region 8 (EPA) Modeling Analysis

CALMET Surface Stations AGC included all the 36 surface stations listed in the MDEQ Draft Protocol ("protocol”) in the CALMET processing.

EPA included all the 36 surface stations listed in the MDEQ Draft Protocol as well as 3 additional surface stations in the CALMET processing.

CALMET Precipitation Stations AGC included 146 precipitation stations in the CALMET processing.

EPA did not include precipitation stations in the CALMET processing.

CALMET Surface Station for Surface Temperature

AGC used the Helena Regional Airport surface station for the surface temperature, as this is the surface station nearest the Montana City plant (7.6 km).

EPA used the Billings Logan Airport surface station for surface temperature. This station is 272.4 km from the Montana City plant.

CALPUFF Puff Splitting AGC included puff splitting, per the protocol.

EPA did not include puff splitting.

CALPUFF Coordinate System AGC used Lambert Conformal Coordinates. The following are the reference coordinates: Reference Latitude: 43.1861N Reference Longitude: -116.2657 W Latitude 1: 43 N Latitude 2: 49 N

EPA used Lambert Conformal Coordinates. The following are the reference coordinates: Reference Latitude: 44.29 N Reference Longitude: -109.5 W Latitude 1: 45 N Latitude 2: 49 N False Easting: 600 meters

CALPUFF Grid Size AGC used a grid size of 2 km. This smaller grid size was selected due to the distance of the closest Class I area to the Montana City plant (Gates of the Mountains, 30 km).

EPA used a grid size of 6 km.

CALPUFF Background Ozone AGC used default background ozone concentrations of 30 parts per billion (ppb) for October through May and 50 ppb for June through September.

EPA used a default background ozone concentration of 80 ppb for the entire year.

CALPOST Monthly Relative Humidity Adjustment Factor

AGC used the monthly relative humidity adjustment factors based on the representative IMPROVE site location for the Class I area, as shown in the MDEQ protocol.

EPA used the monthly relative humidity adjustment factors based on the centroid of the Class 1 Area.

In addition to different modeling methods, AGC also modeled slightly different NOX and SO2 emission rates for the kiln. EPA modeled the Montana City plant based on emissions data that AGC had submitted to MDEQ (and, subsequently, EPA) for the BART applicability analysis. The kiln NOX and SO2 emissions data provided in that submittal was from stack testing performed in April of 2006. In May of 2006, a SO2/NOX analyzer was installed on the kiln exhaust and AGC has collected additional data on SO2 and NOX emissions from the kiln. The data from May 2006 through the end of 2006 show that the maximum actual SO2 and NOx emission rates from the kiln are higher than the SO2 and NOx emission rate originally submitted to MDEQ. The emissions data are summarized in Table 3-3.


TABLE 3-3. KILN EMISSION RATES IN BART DATA SUBMITTAL VS. 2006 ANALYZER DATA

Pollutant

BART Data Submittal to

MDEQ (lb/hr) Comment

2006 Analyzer Data

(lb/hr) Comment SO2 285.83 Stack Test Data,

April 2006 473.87 2006 Maximum Actual 24-

Hour SO2 Emission Rate From Analyzer Data (Hourly

Equivalent) NOX 439.17 Stack Test Data,

April 2006 848.74 2006 Maximum Actual 24-

Hour NOX Emission Rate From Analyzer Data (Hourly

Equivalent) PM10 37.17 Stack Test Data,

April 2006 37.17 Stack Test Data, April 2006

AGC updated the emission rates used in the refined BART applicability modeling to the emission rates based on the analyzer data. Table 3-4 summarizes the emission rates that EPA and AGC modeled for SO2, NOX, and PM10, including the speciated PM10 emissions. The total PM10 emission rates include both the filterable and condensable fractions and are speciated into the following:

▲ Coarse particulate matter (PMC) ▲ Fine particulate matter (PMf) ▲ Sulfates (SO4) ▲ Secondary organic aerosols (SOA) ▲ Elemental carbon (EC)

TABLE 3-4. EXISTING MAXIMUM 24-HOUR SO2, NOX, AND PM10 EMISSIONS (AS HOURLY EQUIVALENTS)

Model Source SO2 NOX Total PM10

SO4 PMc PMf SOA EC

(lb/hr) (lb/hr) (lb/hr) (lb/hr) (lb/hr) (lb/hr) (lb/hr) (lb/hr) EPA Region 8 Kiln 285.83 439.17 37.17 6.80 7.28 21.35 0.93 0.82 Applicability Clinker Cooler 0.00 0.00 6.00 0.00 0.00 5.94 0.00 0.06 AGC Refined Kiln 473.87 848.74 37.17 6.80 7.28 21.35 0.93 0.82 Applicability Clinker Cooler 0.00 0.00 6.00 0.00 0.00 5.94 0.00 0.06

Table 3-5 summarizes the stack parameters that were used to model the kiln and clinker cooler.


TABLE 3-5. SUMMARY OF EXISTING STACK PARAMETERS

Kiln Clinker Cooler Latitude (degrees) 46.544 46.539 Longitude (degrees) -111.921 -111.922 Stack height (ft) 100 50 Stack Diameter (ft) 9 4 Exhaust Velocity (ft/s) 47 54 Exhaust Temperature (K) 384 132

The results of AGC’s refined modeling verified EPA’s BART determination; the results are summarized in Table 3-6. The 98th percentile of the visibility impacts attributable to the kiln and clinker cooler are greater than 0.5 Δdv when compared against a natural background for the Gates of the Mountains Wilderness Area.

TABLE 3-6. EXISTING VISIBILITY IMPAIRMENT ATTRIBUTABLE TO MONTANA CITY PLANT KILN AND CLINKER COOLER

Class I Area

Minimum Distance

98th Percentile Visibility Impact for Each Year

(∆ dv)

Overall 98th

Percentile Visibility Impact

(km) 2001 2002 2003 (dv) Gates of the Mountains 30 2.736 2.874 3.038 2.874


4. SO2 BART EVALUATION

Sulfur, in the form of metallic sulfides (pyrite), sulfate, or organosulfur compounds, is often found in the raw materials used to manufacture cement and in the solid and liquid fuels burned in cement kilns.3 The raw materials and fuels for the Montana City plant are no exception. Sulfur dioxide can be generated by the oxidation of sulfur compounds in the raw materials and fuels during operation of the pyroprocess. Constituents found in fuels, raw materials, and in-process materials, such as the alkali metals (sodium and potassium), calcium carbonate, and calcium oxide often react with SO2 within the pyroprocess to limit emissions of SO2 as much of the sulfur leaves the process in the principle product of the kiln system called clinker. The kiln is the only BART source which emits SO2, thus an SO2 BART evaluation was performed only for the kiln. The maximum actual 24-hour kiln SO2 emission rate that was modeled for the BART applicability determination is summarized in Table 4-1. The SO2 24-hour maximum actual emission rate was determined from analyzer data for 2006.

TABLE 4-1. EXISTING ACTUAL MAXIMUM 24-HOUR SO2 EMISSION RATES

SO2 24-Hour Emission Rate

SO2 Hourly Equivalent Emission Rate

(ton/24-hr) (lb/hr) Kiln 5.69 473.87

4.1 IDENTIFICATION OF AVAILABLE RETROFIT SO2 CONTROL TECHNOLOGIES

Step 1 of the BART determination is the identification of all available retrofit SO2 control technologies. A list of control technologies was obtained by reviewing the U.S. EPA’s Clean Air Technology Center, publicly-available air permits, applications, and technical literature published by the U.S. EPA, state agencies, and Regional Planning Organizations (RPOs). The available retrofit SO2 control technologies are summarized in Table 4-2.

3 Miller, F. MacGregor and Hawkins, Garth J., ”Formation and Emission of Sulfur Dioxide from the Portland Cement

Industry”, Proceedings of the Air and Waste Management Association, June 18-22, 2000.


TABLE 4-2. AVAILABLE SO2 CONTROL TECHNOLOGIES

SO2 Control Technologies

Fuel Substitution Raw Material Substitution Lime Spray Dryer Wet Lime Scrubbing

4.2 ELIMINATE TECHNICALLY INFEASIBLE SO2 CONTROL TECHNOLOGIES Step 2 of the BART determination is to eliminate technically infeasible SO2 control technologies that were identified in Step 1.

4.2.1 FUEL SUBSTITUTION

AGC uses a mixture of coal and petroleum coke as the primary fuels for the kiln; natural gas is combusted during startup. The 2006 fuel usage breakdown, on an energy input basis, was 58 percent petroleum coke, 41 percent coal, and 1 percent natural gas. The sulfur content of the petroleum coke is approximately 4.5 percent and the sulfur content of the coal is approximately 0.8 percent. The design of the long wet kiln system is such that much of the SO2 resulting from fuel combustion is emitted. Therefore, if AGC reduces sulfur in fuel input to the kiln, a corresponding reduction in SO2 emissions from the kiln would be expected. Fuel sulfur content could be reduced by burning a different blend of coal and coke which results in a lower overall sulfur content. Therefore, AGC anticipates that lowering the input of sulfur through fuel substitution would be an effective and technically feasible SO2 control technology for the kiln.

4.2.2 RAW MATERIAL SUBSTITUTION

Sulfide sulfur in the raw materials, usually in the form of iron pyrite, is thermally decomposed and oxidized or “roasted” to form SO2. The pyritic sulfur reacts with oxygen according to the following reaction: 4FeS2 + 11O2 2Fe2O3 + 8SO2 Using raw materials with lower pyritic sulfur content can reduce the potential for SO2 emissions from a wet kiln system. While pyrites are present in the limestone and other raw materials used at the Montana City plant, the concentrations of sulfide sulfur in these materials are already very low. With rare exceptions, cement plants are built at or near a source of limestone, the primary raw material for cement manufacture. To do otherwise is an economic penalty that would cause most plants, including Montana City plant, to be economically infeasible. During the production of cement clinker, the limestone loses about one-third of its weight as CO2. The shipping costs for the “lost” weight in the limestone often can be economically prohibitive.


Since material substitution would result in a negligible reduction in SO2 for the kiln, raw material substitution is not considered an SO2 control technology for the kiln and is removed from further consideration as BART.

4.2.3 LIME SPRAY DRYING

Lime spray drying (LSD) consists of injecting an aqueous Ca(OH)2 suspension in fine droplets into the flue gas. The Ca(OH)2 reacts with SO2 in the flue gas stream to create fine particles of CaSO3 or CaSO4. The moisture evaporates from the particles, and the particles are collected in the particulate matter control device (PMCD) serving the kiln. For optimum effectiveness, the reaction of Ca(OH)2 with SO2 must have adequate gas retention time and must be followed by a PMCD for capturing the sulfates created by the reaction. Unlike a preheater/precalciner kiln system that provides by its design a natural location for a spray dryer type control system to be utilized between the top of the preheater tower and the PMCD, a wet kiln does not have that attribute. In other words, the back-end of a wet kiln has a relatively short retention time between that and it’s associated PMCD. Additionally, the PMCD in place was not designed for the additional loading or for the increased water vapor that would need to be moved through the system to accommodate adequate lime spray drying control. It is also expected that with the added moisture generated from the wet process slurry that rates of evaporation for spray drying to occur could be retarded as compared to other combustion systems that might employ this type of system. Lastly, the added gas cooling that would result from the injection of slurry prior to the PMCD would have the potential for undesirable acid dewpoint conditions to occur in the PMCD that could reduce its effectiveness. For these reasons and as there are no known applications of lime spray dryers on wet cement kilns, this technology considered is technically infeasible and is removed from further consideration as BART.

4.2.4 WET LIME SCRUBBING

Wet lime scrubbing (WLS) is a name for a traditional tailpipe wet scrubber. This process involves passing the flue gas from the main PMCD through a sprayed aqueous suspension of Ca(OH)2 or CaCO3 (limestone) that is contained in an appropriate scrubbing device. In the case of the Montana City plant, the basic underlying economics would dictate the use of ground limestone as the scrubbing reagent. In WLS, the aqueous suspension of scrubbing reagent is not taken to dryness as it is in LSD. The SO2 reacts with the scrubbing reagent to form CaSO3 that is collected and retained as aqueous sludge. Typically, the sludge is dewatered and disposed in an on-site landfill. In some cases involving cement plants, the CaSO3 sludge could be oxidized to CaSO4 and used in the finish mills as a substitute for purchased gypsum for regulation of the setting time of the cement product.


Typically, WLS is considered to have a scrubbing efficiency of up to 90 percent of the SO2 in the flue gas treated by the scrubber.4 WLS is a high maintenance process with high rates of downtime expected as the scrubber matures and corrosion of components becomes a serious problem. Conceivably, a pair of wet scrubbers ultimately would be required so that one is in operation while the other is repaired. Despite these identified drawbacks, WLS is considered a technically feasible BART option.

4.3 RANK OF TECHNICALLY FEASIBLE SO2 CONTROL OPTIONS BY EFFECTIVENESS

The third step in the BART analysis is to rank the technically feasible options according to effectiveness. Table 4-3 presents potential SO2 technically feasible control technologies for the kiln and the associated SO2 emission levels.

TABLE 4-3. RANKING OF TECHNICALLY FEASIBLE KILN SO2 CONTROL TECHNOLOGIES BY EFFECTIVENESS

Pollutant Control Technology

Effectiveness SO2 Emissions Level

(lb/hr)

SO2 Wet Lime Scrubbing 25.35 lb/hr as a 30-day rolling average† Fuel Switching 126.76 lb/hr as a 30-day rolling average *

†Based on a 90% removal efficiency from the 2006 average 24-hour emission rate hourly equivalent (253.52 lb/hr). The 90 percent reduction was applied to the 2006 average 24-hour SO2 emission rate rather than the maximum 24-hour SO2 emission rate to best reflect the performance of the control on a 30 day rolling basis. *Based on a fuel switching scenario that reduces sulfur emissions by 50% from the 2006 average 24-hour emission rate hourly equivalent (253.52 lb/hr). The 90 percent reduction was applied to the 2006 average 24-hour SO2 emission rate rather than the maximum 24-hour SO2 emission rate to best reflect the performance of the control on a 30 day rolling basis.

4.4 EVALUATION OF IMPACTS FOR FEASIBLE SO2 CONTROLS Step four for the BART analysis procedure is the impact analysis. The BART determination guidelines list the four factors to be considered in the impact analysis: • Cost of compliance • Energy impacts • Non-air quality impacts; and • The remaining useful life of the source

4EPA Air Pollution Control Technology Fact Sheet – Flue Gas Desulfurization (FGD) Wet, Spray Dry, and Dry

Scrubbers. A control efficiency of up to 95% is listed. However, at an uptime of 95%, the actual annual control efficiency would be 90.2%


AGC has conducted an impact analysis for the two control options with the highest SO2 control efficiencies: WLS and fuel switching.

4.4.1 WET LIME SCRUBBING

Cost of Compliance AGC obtained a site-specific WLS proposal from a vendor and performed an economic analysis to determine the annualized cost for WLS. AGC divided the annualized cost of WLS by the annual tons of SO2 reduced to determine the cost effectiveness for WLS. The “annual tons reduced” were determined by subtracting the estimated controlled annual emissions from the existing annual emissions. The existing annual emissions are based on the average 24-hour SO2 emission rate (hourly equivalent) in 2006, as recorded by the gas analyzer, multiplied by the 2006 operating hours. The estimated controlled annual emissions were calculated by applying the 90 percent control efficiency to the existing annual emissions. Table 4-4 provides the cost effectiveness analysis related to WLS. Note that the cost effectiveness analysis does not include the cost to construct a new exhaust stack, which would be needed to employ the WLS technology.


TABLE 4-4. COST ANALYSIS FOR WET LIME SCRUBBING

Direct Costs Purchased Equipment Costs Wet Scrubber Unit $5,687,500 Instrumentation (10% of EC) $568,750 Sales Tax (3% of EC) $170,625 Freight (5% of EC) $284,375 Subtotal, Purchased Equipment Cost (PEC) $6,711,250 Direct Installation Costs Foundation (6% of PEC) $402,675 Supports (6% of PEC) $402,675 Handling and Erection (40% of PEC) $2,684,500 Electrical (1% of PEC) $67,113 Piping (30% of PEC) $2,013,375 Insulation for Ductwork (1% of PEC) $67,113 Painting (1% of PEC) $67,113 Subtotal, Direct Installation Cost $5,704,563 Site Preparation N/A Buildings N/A Total Direct Cost $12,415,813 Indirect Costs Engineering (10% of PEC) $671,125 Construction and Field Expense (10% of PEC) $671,125 Contractor Fees (10% of PEC) $671,125 Start-up (1% of PEC) $67,113 Performance Test (1% of PEC) $67,113 Contingencies (3% of PEC) $201,338 Total Indirect Cost $2,348,938 Total Capital Investment (TCI) $14,764,750


TABLE 4-4. COST ANALYSIS FOR WET LIME SCRUBBING (CONTINUED)

Direct Annual Costs Hours per Year (365 days per year, 24 hours per day) 8,760 Operating Labor Operator (0.5 hr/shift, 3 shifts/day, 365 d/yr, $16/hr) $8,760 Supervisor (15% of operator) $1,314 Subtotal, Operating Labor $10,074 Maintenance Labor (0.5 hr/shift, 3 shifts/day, 365 d/yr, $16/hr) $8,760 Material (100% of maintenance labor) $8,760 Subtotal, Maintenance $17,520 Utilities Electricity Pump (kW) 380.48 Cost ($/kW-hr) $0.0537 Subtotal, Electricity $179,082 Limestone Slurry Amount Required (ton/yr) 2,847 Cost ($/ton) $15.00 Subtotal, Lime $42,705 Water Amount Required (gpm) 31.0 Cost ($/1000 gallons) $3.075 Subtotal, Water $50,101 Sludge Disposal Amount Generated (tpy) 5,913 Disposal Fee ($/ton) $23.00 Subtotal, Sludge $135,999 Subtotal, Utilities $407,887 Total Direct Annual Costs $435,481


TABLE 4-4. COST ANALYSIS FOR WET LIME SCRUBBING (CONTINUED)

Indirect Annual Costs Overhead (60% of sum of operating, supervisor, maintenance labor & materials) $153,839 Administrative (2% TCI) $295,295 Property Tax (1% TCI) $147,648 Insurance (1% TCI) $147,648 Capital Recovery (10 year life, 7 percent interest) $2,102,168 Total Indirect Annual Cost $2,846,597 Conclusion Total Annualized Cost $3,282,078

Pollutant Emission Rate Prior to Scrubber (tons SO2/yr) 981 Pollutant Removed (tons SO2/yr) 883 Cost Per Ton of Pollutant Removed $3,716

Energy Impacts A wet scrubber requires an additional fan of considerable horsepower to move the flue gas through the scrubber. Non Air-Quality Impacts WLS may lead to an increase in PM emissions because some particles of limestone or CaSO3 will be entrained in the flue gas and subsequently be emitted from the scrubber. WLS is also known to increase emissions of sulfuric acid mist.5 A frequent steam plume can be expected at the discharge of the wet scrubber that would result in visual impairment in the area. Utilization of a wet scrubber would require the use of a significant amount of water. An appropriately sized wet scrubber would consume approximately 16 million gallons of water per year. Most of this water would be emitted as vapor with a small portion in the sludge that would be generated by the control device. In addition to the consumption of a large amount of water, the WLS technology would also generate a large amount of sludge. Disposal or treatment of WLS sludge presents additional environmental impacts. Remaining Useful Life

5 Innovations in Portland Cement Manufacturing, Portland Cement Association, 2004, pg. 660 & 669


The remaining useful life of the kiln does not impact the annualized cost of WLS because the useful life is anticipated to be at least as long as the capital cost recovery period, which is 10 years.

4.4.2 FUEL SUBSTITUTION

Cost of Compliance The cost of fuel substitution was determined by calculating the cost of the current coal/coke fuel blend and determining the increased cost of switching to combusting a fuel blend that would reduce fuel sulfur content by 50 percent from the 2006 levels. The proposed solution discussed in this evaluation would equate to reducing coke usage from the proportions used in 2006. At this time, AGC has not fully evaluated the potential fuel blends that could be used to reduce fuel sulfur content. The current coal/coke fuel blend costs are based on 2006 fuel usage and cost data for the Montana City plant. The fuel switching costs are based on a switch to an 18.5% coke and 81.5% coal blend, where the specific coal and coke assumed for the blend are the coal and coke that are currently burned at the plant. This fuel blend results in an approximate 50% reduction in fuel sulfur content from the fuel blend used in 2006. Again, AGC has not fully evaluated all potential fuel blends that would result in a 50% reduction in fuel sulfur content; this fuel blend was used only for costing purposes. In practice, AGC may utilize higher quality coal or lower sulfur coke to meet the energy requirements of the kiln. It was assumed in this analysis that fuel switching will not require any capital expenses. The total annual cost of fuel switching was divided by the annual tons of SO2 reduced to determine the cost effectiveness for fuel switching. The “annual tons reduced” were determined by subtracting the estimated controlled annual emissions from the existing annual emissions. The existing annual emissions are based on the average 24-hour SO2 emission rate (hourly equivalent) in 2006, as recorded by the gas analyzer, multiplied by the 2006 operating hours. The estimated controlled annual emission rates were calculated by reducing the existing annual emission rate by 50%. The sulfur content of the existing coal/coke fuel blend is 2.09 percent; this is calculated based on the sulfur contents of the fuels (3 percent for coke and 0.8 percent for coal) and the 2006 fuel usage (by MMBtu). The calculation for the sulfur content of the coal/coke fuel blend is as follows:

905,045 Coke MMBtu 4.5% Sulfur + 643,376 Coal MMBtu 0.8 % Sulfur = 2 96 % Sulfur 1,548,421 Total MMBtu 1,548,421 Total MMBtu

The sulfur content of the fuel switching scenario is calculated based on a 11% coke and 89% coal blend. The calculation of the sulfur content of the fuel blend and the reduced emissions are shown below:

286,458 Coke MMBtu 4.5% Sulfur + 1,261,963 Coal MMBtu 0.8 % Sulfur = 1.48 % Sulfur 1,548,421 Total MMBtu 1,548,421 Total MMBtu


2.96 % Sulfur - 1.48% Sulfur = 50 %

2.96 % Sulfur

253.52 lb 7,741 hr ton 50 % = 491 ton SO2 Reduced hr yr 2,000 lb yr

The cost of fuel switching is summarized in Table 4-5.


TABLE 4-5. SUMMARY OF COST EFFECTIVENESS FOR FUEL SUBSTITUTION

Existing Annual

Emissions

Controlled Annual

Emissions

Reduced Annual

Emissions

Existing Annual Energy

(Coal/Coke) Usage

Existing Coal/Coke

Cost* Average

Heating Value Annual Fuel

Usage Cost Annual

Fuel Cost

Cost of Switching

Fuels Cost

Effectiveness(tons/yr) (tons/yr) (tons/yr) (MMBtu/yr) ($/yr) Btu/lb Tons/yr $/ton ($/yr) ($/yr) ($/ton)

981 492 490 1,548,421 $1,860,967 14,582 (Coke) 9,822 (Coke) 16.38 (Coke) 2,814,066 953,099 1,946.81 8,426 (Coal) 74,885 (Coal) 35.43 (Coal)

*The existing coal/coke cost is based on 2006 actual usage data (31,033 tons coke * $16.38/ton + 38,178 tons Coal * 35.43/ton = $1,860,967).


Energy Impacts and Non Air-Quality Impacts There are no energy or non-air quality impacts associated with fuel switching. Remaining Useful Life The remaining useful life of the kiln does not impact the annualized costs for fuel switching, since, for this analysis, it is assumed that fuel switching will not require any capital costs.

4.5 EVALUATION OF VISIBILITY IMPACT OF FEASIBLE SO2 CONTROLS A final impact analysis was conducted to assess the visibility improvement for existing emission rates when compared to the emission rates of WLS and fuel switching. The existing emission rates and emission rates associated with WLS and fuel switching were modeled using CALPUFF. The existing emission rates are the same rates that were modeled for the BART applicability analysis. The SO2 emissions rates associated with WLS and fuel switching are the proposed BART emission limits in lb/hr based on reductions from the 2006 average 24-hour emission rate (as an hourly equivalent). The SO2 emission reductions from the WLS and fuel switching control options were applied to the 2006 average 24-hour SO2 emission rate rather than the maximum 24-hour SO2 emission rate because the BART limit is proposed as a 30-day rolling average. Had the reductions been applied to the maximum 24-hour SO2 emission rates, the controlled emission rates would be much higher than what AGC anticipates could be achieved by the WLS and fuel switching control options on a 30-day rolling basis. The emission rates are summarized in Table 4-6.

TABLE 4-6. SUMMARY OF EMISSION RATES MODELED IN SO2 CONTROL VISIBILITY IMPACT ANALYSIS

Emission Rate Scenario Emission Rate SO2 NOX PM10 (lb/hr) (lb/hr) (lb/hr) WLS 25.35 848.74 37.17 Fuel Substitution 126.76 848.74 37.17

Comparisons of the existing visibility impacts and the visibility impacts based on WLS and fuel switching for the Gates of the Mountains Wilderness Area are provided in Table 4-7. The visibility improvement associated with WLS and fuel switching are also shown in Table 4-7; this value was calculated as the difference between the existing visibility impairment and the visibility impairment for the controlled emission rates as measured by the 98th percentile modeled visibility impact.


TABLE 4-7. SUMMARY OF MODELED IMPACTS FROM SO2 CONTROL VISIBILITY IMPACT ANALYSIS

98% Impact

(Δdv) Improvement Existing 2.87 --

Fuel Switching 2.70 6.05% WLS 2.63 8.39%

As shown in Table 4-7, the installation of a WLS on the kiln results in an 8.39 percent improvement to the existing visibility impairment. Fuel switching results in a 6.05 percent improvement to the existing visibility impairment. Therefore, utilization of WLS as compared to fuel switching results in only a 2% incremental improvement (0.07 Δdv). The minimal visibility improvement was expected due to the low contribution of sulfates to the existing visibility impairment when compared to nitrates.

4.6 PROPOSED BART FOR SO2

In order to determine BART for SO2, AGC evaluated each control option’s cost of compliance, energy impacts, and non-air quality impacts, as well as the remaining useful life of the kiln. Table 4-8 summarizes the cost effectiveness for each control option based on the tons of SO2 reduced and the visibility improvement in deciviews. The cost effectiveness for the fuel switching is $1,947 per ton of SO2 reduced and $5.5 million per deciview of visibility improvement. This corresponds to a nominal visibility improvement from 2.87 Δdv to 2.70 Δdv. The cost effectiveness for the WLS is $3,716 per ton of SO2 reduced and $13.6 million per deciview of visibility improvement. This corresponds to a nominal visibility improvement from 2.87 Δdv to 2.63 Δdv. The incremental cost of utilizing WLS as opposed to fuel switching is $33,271,129 per deciview.

TABLE 4-8. SUMMARY OF COST EFFECTIVENESS OF SO2 CONTROL TECHNOLOGIES

Existing

Emissions Controlled Emissions

Reduced Annual

Emissions

Annual Cost Cost Effectiveness

(tons/yr) (tons/yr) (tons/yr) ($/yr) ($/ton) Fuel Switching 981 492 490 953,099 1,947

WLS 981 98 883 3,282,078 3,716

Base 98th Percentile

Impact

98th Percentile Impact

98th Percentile

Improvement

98th Percentile

Improvement

Cost Effectiveness

(DV) (DV) (DV) (%) ($/DV) Fuel Switching 2.87 2.70 0.17 6.05 5,477,581

WLS 2.87 2.63 0.24 8.39 13,618,583


Based on the five step analysis outlined by EPA, fuel switching and WLS were identified as the two technically feasible technologies. Cost, energy and environmental impacts were assessed for both technologies and the visibility improvements associated with both options were evaluated against existing conditions. This analysis demonstrates that the cost of compliance associated with both control options is high while the visibility impact analysis demonstrates that the visibility improvements associated with both control options are nominal due to the fact that the percentage of visibility impairment attributable to SO4 is relatively low. As a result, AGC has determined that additional SO2 control technologies (fuel switching and WLS) would provide little visibility improvement and require significant expenditures. Therefore, AGC proposes that limiting the kiln to the existing levels of SO2 emissions constitutes BART for the kiln.


5. NOX BART EVALUATION

In Portland cement kilns, the NOx that is generated is primarily classified into one of two categories, i.e., thermal NOx or fuel NOx6. Thermal NOx occurs as a result of the high-temperature oxidation of molecular nitrogen present in the combustion air. Fuel NOx is created by the oxidation of nitrogenous compounds present in the fuel. It is also possible for nitrogenous compounds to be present in the raw material feed and become oxidized to form additional NOx referred to as feed NOx. Due to the high flame temperature in the burning zone of the rotary kiln (3400o F), NOx emissions from the kiln tend to be mainly comprised of thermal NOx. Although NOx emissions from cement kilns include both nitrogen oxide (NO) and nitrogen dioxide (NO2), typically, less than 10% of the total NOx in the flue gas is NO2.7 The kiln is the only BART source which emits NOx, thus a NOx BART evaluation was performed only for the kiln. The maximum actual 24-hour kiln NOx emission rate that was modeled for the BART applicability determination is summarized in Table 4-1. The NOx 24-hour maximum actual emission rate was determined from analyzer data for 2006.

TABLE 5-1. EXISTING ACTUAL MAXIMUM 24-HOUR NOX EMISSION RATES

NOx 24-Hour Emission Rate

NOx Hourly Equivalent Emission Rate

(ton/24-hr) (lb/hr) Kiln 10.18 848.74

5.1 IDENTIFICATION OF AVAILABLE RETROFIT NOX CONTROL TECHNOLOGIES

Step 1 of the BART determination is the identification of all available retrofit NOX control technologies. A list of control technologies was obtained by reviewing the U.S. EPA’s Clean Air Technology Center, control equipment vendor information, publicly-available air permits, applications, and technical literature published by the U.S. EPA and the RPOs. The available retrofit NOX control technologies are summarized in Table 5-2.

6 NOx Formation and Variability in Portland Cement Kiln Systems, Penta Engineering, December 1998. 7 IBID.


TABLE 5-2. POSSIBLE NOX CONTROL TECHNOLOGIES

Kiln Control Technologies

Low NOx Burner Flue Gas Recirculation CKD Insufflation Mid-Kiln Firing of Tires Selective Noncatalytic Reduction Selective Catalytic Reduction

5.2 ELIMINATE TECHNICALLY INFEASIBLE NOX CONTROL TECHNOLOGIES Step 2 of the BART determination is to eliminate technically infeasible NOX control technologies that were identified in Step 1.

5.2.1 LOW-NOX BURNER IN THE ROTARY KILN

Low NOx burners (LNBs) reduce the amount of NOx formed at the flame. The principle of all LNBs is the same: stepwise or staged combustion and localized exhaust gas recirculation (i.e. at the flame). As applied to the rotary cement kiln, the low-NOx burner creates primary and secondary combustion zones at the end of the main burner pipe to reduce the amount of NOx initially formed at the flame. In the high-temperature primary zone, combustion is initiated in a fuel-rich environment in the presence of a less than stoichiometric oxygen concentration. The oxygen-deficient condition at the primary combustion site minimizes thermal and fuel NOx formation and produces free radicals that chemically reduce some of the NOx that is being generated in the flame. In the secondary zone, combustion is completed in an oxygen-rich environment. The temperature in the secondary combustion zone is much lower than in the first; therefore, lower NOx formation is achieved as combustion is completed. CO that has been generated in the primary combustion zone as an artifact of the sub-stoichiometric combustion is fully oxidized in the secondary combustion zone. The EPA has indicated that a 14% reduction in NOx emissions may be anticipated in switching from a direct-fired standard burner to an indirect-fired LNB8. This is based on a study conducted on an indirect-fired LNB at the Dragon Product Company cement kiln at the plant located in Thomaston, Maine. However, the EPA has also determined that the [emission reduction] contribution of the LNB itself and of the firing system conversion [direct to indirect] can not be isolated from the limited data available9. The terms direct and indirect firing have unique meaning in the context of kiln firing (unlike the more general meanings where direct firing implies that the products of combustion contact the

8 NOX Control Technologies for the Cement Industry, EC/R Incorporated, Chapel Hill, NC, USA, U.S. EPA

Contract NO. 68-D98-025, U.S. EPA RTP, September 19, 2000. 9 USEPA, Office of Air Quality Planning and Standards. Alternative Controls Technology Document - NOx

Emissions from Cement manufacturing. EPA-453/R-94-004, Page 5-5 to 5-8.


process materials whereas indirect firing involves a heat transfer medium). In kiln firing, direct and indirect firing describes the manner in which pulverized fuel is conveyed from the fuel grinding mill to the burner. In the direct firing configuration, fuel is pneumatically conveyed directly from the coal mill to the burner. The quantity of air introduced to the primary combustion zone is dictated by the minimum air requirements of the coal mill and the conveyance system, rather than the optimum flame requirements. The Montana City plant kiln uses a direct firing system. In the indirect firing configuration, the coal mill air is separated from the pulverized fuel which is stored in a tank before being fed to the kiln. The pulverized fuel is then conveyed to the burner with the quantity of air that is optimum for flame considerations. There have been no controlled studies conducted on cement kilns that verify that this method of burning solid fuel reduces the formation of NOx. The AGC Midlothian, Texas plant, which also operates direct-fired wet kilns, utilizes a direct fired LNB system that consists of a plugged annual burner pipe. In this design, the burner pipe has a central plug, which reduces the pressure at the core of the jet. As a result, the pressure of the primary air jet is relieved inward, reducing the rate of the expansion of the flame. This produces a non-divergent flame that minimizes surface area of the flame and maintains the fuel concentrated in the core of the flame. The annual burner pipe is shown in Figures 5-1 and 5-2.

FIGURE 5-1 ANNUAL BURNER PIPE WITH CONTRACTED FLAME

FIGURE 5-2 SCHEMATIC OF ANNUAL BURNER PIPE

When compared to simple free jet burners without the annual nozzle, these burners enhance NOX control by reducing flame turbulence, delaying fuel/air mixing, and


establishing a fuel rich core in the flame for initial combustion. Low-NOx burners are considered to be a technically feasible option for NOX control.

5.2.2 FLUE GAS RECIRCULATION

Flue gas recirculation involves the use of oxygen-deficient flue gas from some point in the process as a substitute for primary air in the main burner pipe in the rotary kiln. Flue gas recirculation (FGR) lowers the peak flame temperature and develops localized reducing conditions in the burning zone through a significant reduction of the oxygen content of the primary combustion “air.” The intended effect of the lower flame temperature and reducing conditions in the flame is to decrease both thermal and fuel NOx formation in the rotary kiln. While FGR is a practiced control technology in the electric utility industry, AGC is not aware of any attempt to apply FGR to a cement kiln because of the unique process requirements of the industry, i.e., a hot flame is required to complete the chemical reactions that form clinker minerals from the raw materials. The process of producing clinker in a cement kiln requires the heating of raw materials to about 2700°F for a brief but appropriate time to allow the desired chemical reactions that form the clinker minerals to occur. A short, high-temperature flame of about 3400°F is necessary to meet this process requirement. The long/lazy flame that would be produced by FGR would result in the production of lower or unacceptable quality clinker because of the resulting undesirable mineralogy. Clinkering reactions must take place in an oxidizing atmosphere in the burning zone to generate clinker that can be used to produce acceptable cement. FGR would tend to produce localized or general reducing conditions that also could detrimentally affect clinker quality. Due to these important limitations on the application of FGR and the lack of a successful demonstration on a cement kiln in the United States, FGR is not a technically feasible control option for NOx control at this time.

5.2.3 CEMENT KILN DUST INSUFFLATION

Cement kiln dust (CKD) is a residual byproduct that can be produced by any of the four basic types of cement kiln systems. CKD is most often treated as a waste even though there are some beneficial uses. However, as a means of recycling usable CKD to the cement pyroprocess, CKD sometimes is injected or insufflated into the burning zone of the rotary kiln in or near the main flame. The presence of these cold solids within or in close proximity to the flame has the effect of cooling the flame and/or the burning zone thereby reducing the formation of thermal NOx. The insufflation process is somewhat counterintuitive because a basic requirement of a cement kiln is a very hot flame to heat the clinkering raw materials to about 2700o F in as short a time as possible. Because there is an increased requirement for thermal energy in the burning zone when insufflation is employed, it is not an attractive technology for recirculation of CKD in wet kiln systems. Other, more efficient procedures are available. Therefore, this option is removed from consideration for BART.


5.2.4 MID-KILN FIRING OF SOLID FUEL (TIRES) WITH MIXING AIR FAN

Secondary combustion is defined as follows: a portion of the fuel is fired in a location other than the burning zone. This reduces thermal NOx generation because the temperature in the secondary combustion zone is less than 2100 °F. Mid-kiln firing (MKF) of solid fuels, such as used tires, is an example of secondary combustion. MKF allows part of the kiln fuel to be burned at a material calcination temperature (secondary combustion zone) which is much lower than the clinker burning temperature. The Cadence feed form MKF technology was first introduced in 1989. It is comprised of three primary components: (1) a staging arm or “feed fork,” that picks up the fuel modules and positions them for entry into the kiln, (2) two pivoting doors that open to allow the fuel to drop into the kiln, and (3) a drop tube that extends through the side wall of the kiln. In addition to these basic components, feed fork technology also requires a delivery system which positions the fuel models so they can be picked up by the feed fork and a mechanism for opening the doors so the fuel can enter the kiln. Due to rotation of the kiln, fuel can only be injected once per revolution from the top, as shown in Figure 5-3.

FIGURE 5-3. MID-KILN FIRING SCHEMATIC10

High-pressure air, in the range of a 2-10 percent replacement of the primary combustion air, could be injected through the shell of the rotary kiln and into the calcining zone to where a mixing air fan mixes the air with the gas and fuel within the rotary kiln for more complete combustion of the solid fuel. By adding fuel mid-kiln, MKF changes both the flame temperature and flame length. These changes should reduce thermal NOX formation by burning part of the fuel at a lower temperature and by creating reducing conditions at the mid-kiln fuel injection point which may destroy some of the NOX formed upstream in the kiln burning zone.

10 NOX Control Technologies for the Cement Industry, EC/R Incorporated, Chapel Hill, NC, USA, U.S. EPA

Contract NO. 68-D98-025, U.S. EPA RTP, September 19, 2000.


The discontinuous fuel feed from MKF may also result in increased carbon monoxide (CO) emissions. To control CO emissions, the kiln may have to have increased combustion air which can decrease production capacity. AGC currently utilizes MKF of tires on three wet kilns at the Midlothian, Texas plant; approximately 4 million tires are burned each year for the three wet kilns. This accounts for about 20% of the fuel usage (BTU basis) at the Midlothian plant. It is estimated that approximately 1.3 million tires would be required at Montana City to achieve the same fuel replacement and equivalent NOX reductions. A study by the Montana Environmental Quality Council11 estimated that between 527,400 and 879,000 waste tires are generated each year in Montana. Therefore, in order for the Montana City plant to obtain the required amount of tires, tires would likely need to be imported from surrounding states. Furthermore, the Holcim Trident cement plant located in Three Forks, Montana, is currently seeking approval from MDEQ to burn up to 1,137,539 tires per year as supplemental fuel. In the Draft Environmental Impact Statement submitted by Holcim to MDEQ in July 2006, Holcim states that it will transport tires from out of state to supplement the number of tires available in Montana. This would create an additional strain on the supply of tires available in the region for the Montana City plant. Transportation of tires from surrounding states would increase the cost associated with MKF and generate additional air pollutant emissions for motor vehicle transportation and fugitive dust from traffic at the Montana City plant. Therefore, while MKF of tires is technically feasible, AGC proposes to eliminate this control option from further consideration as BART due to the supply shortage of tires.

5.2.5 SELECTIVE NONCATALYTIC REDUCTION

In the relatively narrow temperature window of 1600 to 1995°F, ammonia (NH3) reacts with NOx without the need for a catalyst to form water and molecular nitrogen in accordance with the following simplified reactions. 4NO + 4NH3 + O2 4N2 + 6H2O 2NO2 + 4NH3 + O2 3N2 + 6H2O As applied to NOx control from cement kilns and other combustion sources, this technology is called selective noncatalytic reduction (SNCR). Above this temperature range, the NH3 is oxidized to NOx thereby increasing NOx emissions. Below this temperature range, the reaction rate is too slow for completion and unreacted NH3 may be emitted from the pyroprocess. This temperature window generally is available at some location within the rotary kiln. The NH3 could be delivered to the kiln shell through the use of anhydrous NH3, or an aqueous solution of NH3 (ammonium hydroxide) or urea.

11 Status of and Alternatives for Management of Waste Tires in Montana: Report to the 56th Legislature,

Montana Environmental Quality Council, October 1998.


A concern about application of SNCR technology is the breakthrough of unreacted NH3 as “ammonia slip” and its subsequent reaction in the atmosphere with SO2, sulfur trioxide (SO3), hydrogen chloride (HCl) and/or chlorine (Cl2) to form a detached plume of PM10 –PM2.5. SNCR is currently being used successfully as an independent technology on wet cement kiln systems in Europe and recently, authorization was granted to AGC to test SNCR at its wet kiln in Midlothian, Texas. AGC installed a full scale SNCR system on one of its wet kilns and the system has been running for several months; it is achieving a 35 to 40% NOx reduction on a consistent basis. As SNCR is currently being operated on one of AGC’s wet cement kiln at the Midlothian Texas plant, it is considered a technically feasible NOx control option for the Montana City plant

5.2.6 SELECTIVE CATALYTIC REDUCTION

Selective Catalytic Reduction (SCR) is an add-on control technology for the control of emissions of the oxides of nitrogen (NOx) from a combustion process. SCR has been successfully employed in the electric power industry. The basic SCR system consists of a system of catalyst grids placed in series with each other within a vessel that is located in a part of the process where the normal flue gas temperature is in the required range. An ammonia-containing reagent is injected at a controlled rate upstream of the catalyst grids that are designed to ensure relatively even flue gas distribution within the grids, to provide good mixing of the reagent and flue gas, and to result in minimum ammonia (NH3) slip.12 The NH3 reacts with NOx compounds (i.e., NO and NO2) on the surface of the catalyst in equal molar amounts (i.e., one molecule of NH3 reacts with one molecule of NOx). Common reagents include aqueous NH3, anhydrous NH3 and urea [(NH2)2CO]. In the presence of the catalyst, the injected ammonia is converted by OH- radicals to ammonia radicals (i.e., NH2

-), which, in turn, react with NOx to form N2 and H2O. The SCR catalyst enables the necessary reactions to occur at lower temperatures than those required for Selective Non-Catalytic Reduction (SNCR). While catalysts can be effective over a larger range of temperatures, the optimal temperature range for SCR is 570 - 750º F. The catalyst system used in SCR applications usually consists of (1) a porous honeycomb of a ceramic substrate onto which catalyst has been attached to the surface of the ceramic material, or (2) a flat or corrugated plate onto which catalytic material has been deposited on the surface. A porous metal oxide with a high surface area-to-volume ratio acts as a catalyst base. On this base, typically titanium dioxide (TiO2), one or more metal oxide catalysts are deposited in various concentrations. In SCR applications, the active catalyst material typically consists of vanadium pentoxide (V2O5), tungsten trioxide (WO3), and molybdenum trioxide (MoO3) in various combinations. The composition, also known as the catalyst formulation, is tailored by the catalyst vendor to best suit a particular SCR

12 Slip refers to the quantity of unreacted reagent that exits the SCR reactor.


application. Catalyst deactivation through poisoning, fouling, masking, sintering and erosion are common problems for SCR catalysts that, without careful process design and operation, could be exacerbated. If not fouled by sulfur dioxide (SO2), the catalysts used in SCR have a propensity to oxidize sulfur dioxide (SO2) in the flue gas to sulfur trioxide (SO3), a more undesirable pollutant. Because the reaction rate of NH3 and NOx is temperature dependent, the temperature of the flue gas stream to be controlled is the most important consideration in applying SCR technology to any combustion source. The optimum temperature range for SCR application is about 300º C (570º F) to 450º C (840º F). This range of normal process temperature may be found in the exhaust gas from the wet kiln at the PMCD inlet. SCR has not been applied to a cement plant of any type in the United States. SCR has been applied successfully at a cement plant in Solnhofen, Germany. The Solnhofen plant has a kiln with a preheater tower as opposed to the wet kiln system at AGC’s Montana City plant. SCR has also been successfully applied at a cement plant in Moncelice, Italy; however, this plant is also a preheater plant as opposed to the wet cement kiln system at Montana City. Earlier this year, as part of permitting a new cement plant to be located on the Moapa Pauite Indian Reservation in Nevada (Moapa Paiute plant), AGC carefully assessed all of the publicly available information regarding SCR application at Solnhofen to determine (1) whether the Solnhofen testing indicates that the SCR technology exceeds the performance of SNCR and (2) whether the technology is commercially available for preheater/precalciner system such as that intended for AGC’s proposed Moapa Paiute plant. In order to ensure a comprehensive review, AGC engaged an independent expert in SCR technology to conduct an extensive analysis of SCR and its availability in relation to the Moapa Paiute plant. This analysis determined that the SCR system at Solnhofen does not result in a lower NOx emission rate than that which can be gained from SNCR.13 The study also concluded that the cost and time required for pilot testing to select the appropriate catalyst and SCR size/configuration needed to achieve the same emission levels achievable by SNCR is unknown. In the draft Moapa Paiute PSD permit put out for public comment this spring, EPA Region 9 concluded that SCR did not constitute BACT for cement kilns as it could not be demonstrated to outperform SNCR (that permit is expected to be issued upon completion of the Endangered Species Act consultation). The major SCR vendors have also indicated that SCR is not commercially available for cement kilns at this time. The St. Lawrence Cement Company recently issued a Request for Proposals (RFP) for SCR for the company’s proposed new cement kiln in Greenport, New York. Of the four major vendors contacted, two, Lurgi PSI Inc. (Lurgi) and Babcock & Wilcox, did not provide any proposal, with Lurgi stating that their technology was not yet ready for commercial release. A third with relevant experience from the Solnhofen

13 Schreiber, Robert J., Evaluation of Suitability of Selective Catalytic Reduction for Use in Portland Cement

Manufacturing, 2006


demonstration plant, KWH,14 indicated that technical uncertainties prevented them from designing an SCR system without pilot plant testing. Only Alstom provided a proposal that suggested SCR could be supplied to a cement kiln system. However, careful review of the Alstom proposal, indicated that the Alstom proposal did not identify a commercial SCR system that would be viable for a cement kiln system application AGC has reviewed the publicly available SCR assessments and vendor documents related to the Greenport plant. Ash Grove believes that the Greenport vendor evaluation continues to be relevant and supports the conclusion that an SCR system is not commercially available as defined in the NSR Workshop Manual, pages B.17 and B.18, which states that:

…Two key concepts are important in determining whether an undemonstrated technology is feasible: “availability” and “applicability.“

As explained in more detail below, a technology is considered "available" if it can be obtained by the applicant through commercial channels or is otherwise available within the common sense meaning of the term. An available technology is “applicable” if it can reasonably be installed and operated on the source type under consideration. A technology that is available and applicable is technically feasible. Availability in this context is further explained using the following process commonly used for bringing a control technology concept to reality as a commercial product: concept stage; research and patenting; bench scale or laboratory testing; pilot scale testing; licensing and commercial demonstration; and commercial sales. A control technique is considered available, within the context presented above, if it has reached the licensing and commercial sales stage of development. A source would not be required to experience extended time delays or resource penalties to allow research to be conducted on a new technique. Neither is it expected that an applicant would be required to experience extended trials to learn how to apply a technology on a totally new and dissimilar source type. Consequently, technologies in the pilot scale testing stages of development would not be considered available. An exception would be if the technology were proposed and permitted under the qualifications of an innovative control device consistent with the provisions of 40 CFR 52.21(v) or, where appropriate, the applicable SIP [in which case it would be considered available]. Commercial availability by itself, however, is not necessarily sufficient basis for concluding a technology to be applicable and therefore technically feasible. Technical feasibility, as determined in Step 2, also means a control option may reasonably be deployed on or "applicable" to the source type under consideration. (NSR Page B.18) As SCR would require pilot scale testing, Ash Grove concluded that SCR was not “available” with respect to the Moapa Paiute plant because it was not commercially

14 KWH teamed with Elex, a German engineering firm who was responsible for some aspects of the Solnhofen

installation. Elex holds a patent covering certain applications of SCR to cement kilns in the United States.


available. This determination has also been the finding of the Florida DEP as recently as March 2005, which in a BACT determination for Florida Rock Industries, Newberry Plant, concluded that “there has been no pilot study conducted in the United States, and there have been no indications that a pilot plant will be constructed to test SCR by any Portland cement facilities in the United States.” Further, “…Some additional time would be needed to conduct tests to determine the correct catalyst formulation” and, “The Department does not consider SCR necessary to achieve a BACT level of control in Florida.”15.

In conclusion, AGC has determined that SNCR is as good as SCR based on Solnhofen and, due to commercial unavailability that AGC determined for the Moapa Paiute plant and since there is no known application of SCR on a wet kiln system, SCR is eliminated from further consideration as BART for NOx control at the Montana City plant.

5.3 RANK OF TECHNICALLY FEASIBLE NOX CONTROL OPTIONS BY EFFECTIVENESS

The third step in the BART analysis is to rank the technically feasible options according to effectiveness. Table 5-3 presents potential NOx technically feasible control technologies for the kiln and the associated NOx emission levels. The emission rates are largely based AGC’s experience with wet kiln NOX control technologies at AGC’s Midlothian, Texas plant.

15 Air Permit No: 0010087-013-AC; PSD-FL-350 - BACT Determination, Comment written by Al Linero,

Florida DEP, 3/30/2005, Pages BD-10 and BD-11.


TABLE 5-3. RANKING OF TECHNICALLY FEASIBLE KILN NOX CONTROL TECHNOLOGIES BY EFFECTIVENESS

Pollutant Control Technology Effectiveness NOX Emissions Level

(lb/hr)

NOX LNB and SNCR 227.25 lb/hr (~35 % control)† SNCR 295.36 lb/ hr (~35 % control) § LNB 422.59 lb/ hr (~7% control) ††

†The effectiveness level for SNCR and LNB corresponds to a 50% NOX reduction from the 2006 average 24-hour emission rate as an hourly equivalent (454.50 lb/hr). The 90 percent reduction was applied to the 2006 average 24-hour NOx emission rate rather than the maximum 24-hour NOx emission rate to best reflect the performance of the control on a 30 day rolling basis. § The effectiveness level for SNCR corresponds to a 35% NOX reduction from the 2006 average 24-hour emission rate as an hourly equivalent (454.50 lb/hr). The 90 percent reduction was applied to the 2006 average 24-hour NOx emission rate rather than the maximum 24-hour NOx emission rate to best reflect the performance of the control on a 30 day rolling basis. ††The effectiveness level for the direct-fired LNB system is based on AGC’s experience with a system at the Midlothian, Texas plant. The effectiveness corresponds to a 7% NOX reduction from the 2006 average 24-hour emission rate as an hourly equivalent (454.50 lb/hr). The 90 percent reduction was applied to the 2006 average 24-hour NOx emission rate rather than the maximum 24-hour NOx emission rate to best reflect the performance of the control on a 30 day rolling basis.

5.4 EVALUATION OF IMPACTS FOR FEASIBLE NOX CONTROLS Step four for the BART analysis procedure is the impact analysis. The BART determination guidelines list four factors to be considered in the impact analysis:

• Cost of compliance • Energy impacts • Non-air quality impacts; and • The remaining useful life of the source

5.4.1 SNCR AND LNB

Cost of Compliance Since AGC is proposing the most stringent control option as BART, the cost of compliance is not evaluated. Energy Impacts and Non Air-Quality Impacts SNCR systems require electricity to operate the blowers and pumps. The generation of the electricity will most likely involve fuel combustion, which will cause emissions. While the required electricity will result in the emissions, the emissions should be small compared to the reduction in NOx that would be gained by operating an SNCR system


Ammonia slip from SNCR systems occurs either from ammonia injection at temperatures too low for effective reaction with NOx, leading to an excess of unreacted ammonia, or from over-injection of reagent leading to uneven distribution; which also leads to an excess of unreacted ammonia. Based on AGC’s experience at the Midlothian, Texas plant, we believe that ammonia slip will be less than 10 ppm above baseline emissions. While the presence of ammonia slip is recognized here as an impact attributable to SNCR systems, it is an air-quality impact and so is legally not part of the BART analysis process. Remaining Useful Life The remaining useful life of the kiln does not impact the annualized costs of SNCR because the useful life is anticipated to be at least as long as the capital cost recovery period, which is 10 years.

5.5 EVALUATION OF VISIBILITY IMPACT OF FEASIBLE NOX CONTROLS The final impact analysis was conducted to assess the visibility improvement for existing emission rates when compared to the emission rates of the SNCR and LNB combined control option. The existing emission rates and emission rates associated with SNCR and LNB were modeled using CALPUFF. The existing emission rates are the same rates that were modeled for the BART applicability analysis. The NOx emission rate associated with the SNCR and LNB control option is the 2006 average 24-hour NOx emission rate (hourly equivalent) reduced by 50 percent. This 50 percent reduction was applied to the 2006 average 24-hour NOx emission rate rather than the maximum 24-hour NOx emission rate because the BART limit is proposed as a 30-day rolling average. Had the 50 percent reduction been applied to the maximum 24-hour NOx emission rate, the controlled emission rate would be much higher than what AGC anticipates could be achieved by the SNCR and LNB on a 30-day rolling basis. The emission rates are summarized in Table 5-4.


TABLE 5-4. SUMMARY OF EMISSION RATES MODELED IN NOX CONTROL VISIBILITY IMPACT ANALYSIS

Unit Emission Rate Scenario Emission Rate SO2 NOX PM10 (lb/hr) (lb/hr) (lb/hr) Kiln SNCR and LNB 473.87 227.25 37.17 Existing 473.87 848.74 37.17

Comparisons of the 98th percentile existing visibility impacts and the visibility impacts based on LNB and SNCR for the Gates of the Mountains Wilderness Area are provided in Table 5-5. The visibility improvement associated with LNB and SNCR are also shown in Table 5-5; this was calculated as the difference between the existing visibility impairment and the visibility impairment for the remaining control options as measured by the 98th percentile modeled visibility impact.

TABLE 5-5. NOX CONTROL VISIBILITY IMPACT ANALYSIS

98% Impact

(Δdv) Improvement Existing 2.874 -

SNCR and LNB 1.377 52.09% As seen in Tables 5-5, the SNCR and LNB result in a visibility improvement of 52.09 percent.

5.6 PROPOSED BART FOR NOX

Based on the five step analysis outlined by EPA, SNCR with LNB was identified as the sole technically feasible add-on control technology. Cost, energy and environmental impacts were assessed for this technology and the visibility improvements were evaluated against existing conditions. Consistent with EPA guidance, economic impacts were not assessed as AGC was willing to utilize the highest ranked control technology. The visibility impact analysis demonstrates that the utilization of SNCR and LNB to achieve a 227.25 lb/hr NOx emission rate results in significant visibility improvements. Neither non-air quality nor energy impacts associated with this control technology are material and so do not present a basis for eliminating SNCR/LNB in favor of retaining the existing rates as BART. Therefore, AGC proposes that a direct-fired LNB system with SNCR is BART for NOx. AGC proposes to comply with a BART emissions limit of 227.25 lb/hr on a 30-day rolling basis. Compliance with the emission limit will be demonstrated by continuous emissions monitoring.


6. PM BART EVALUATION

PM is generated by the kiln and clinker cooler. The PM emissions are from the kiln are currently controlled by an ESP, and the PM emissions from the clinker cooler are controlled by a baghouse. The maximum daily PM10 emission rates that were modeled for the BART applicability determination are summarized in Table 6-1.

TABLE 6-1. EXISTING MAXIMUM 24-HOUR PM10 EMISSION RATE

PM10 24-Hour Emission Rate

PM10 Hourly Emission Rate

(ton/24-hr) (lb/hr) Kiln 0.45 37.17 Clinker Cooler 0.07 6.00

A comparison of Table 6-1 with Table 4-1 and Table 5-1 shows that the current PM10 emission rates for the kiln and clinker cooler are much less than the current emission rates of SO2 and NOX for the kiln. The low PM10 emission rates correspond to low visibility impacts attributable to PM10 when compared to the impacts attributable to SO2 and NOX, as shown in Table 6-2.

TABLE 6-2. VAP VISIBILITY IMPAIRMENT CONTRIBUTIONS AT GATES OF THE MOUNTAINS

98th Percentile Impact

Visibility Impairment Attributable to SO4

1 Visibility Impairment Attributable to NO3

2 Visibility Impairment Attributable to PM10

3

(∆ dv) (%) (%) (%) 2.16 23.1 66.5 10.4

1 The visibility impairment attributable to SO4 is primarily from SO2 emissions. A very small portion is from SO4 emitted as condensable particulate. 2 The visibility impairment attributable to NO3 is entirely from NOX emissions. 3 The visibility impairment attributable to PM10 is the sum of the visibility impairment attributable to all modeled primary PM species (PMc, PMf, EC, and SOA).

As mentioned, the kiln has an existing ESP for particulate matter control. The ESP is the most effective particulate matter control device for a wet kiln, due to the temperature of the exhaust exiting the kiln. The exhaust from the kiln is well over 5000F at times under normal operations and would require an extensive gas conditioning system to operate a fabric filter PMCD properly. Also, as mentioned, the clinker cooler has an existing baghouse for particulate matter control. A baghouse is currently the best device for controlling particulate matter from a source. As no particulate control devices were identified that are more effective than the existing PMCDs, AGC proposes that no additional PM control technologies are required for either the kiln or clinker cooler for BART. AGC believes that the existing controls are the best, most technically feasible controls for these types of sources. Because AGC is proposing to retain the most effective particulate


control devices on the two BART units, there is no need to evaluate other impacts in establishing these control technologies as BART.

Ash Grove Cement Company | Regional Haze 2nd Implementation Period Four-Factor Analysis B-1 Trinity Consultants

APPENDIX B: 2011 RESPONSE

regional haze 2nd implementation period four-factor …

Documents